Daily arXiv

In this work, we recover the underlying 3D structure of non-geometrically consistent scenes. We focus our analysis on hand-drawn images from cartoons and anime. Many cartoons are created by artists without a 3D rendering engine, which means that any new image of a scene is hand-drawn. The hand-drawn images are usually faithful representations of the world, but only in a qualitative sense, since it is difficult for humans to draw multiple perspectives of an object or scene 3D consistently. Nevertheless, people can easily perceive 3D scenes from inconsistent inputs! In this work, we correct for 2D drawing inconsistencies to recover a plausible 3D structure such that the newly warped drawings are consistent with each other. Our pipeline consists of a user-friendly annotation tool, camera pose estimation, and image deformation to recover a dense structure. Our method warps images to obey a perspective camera model, enabling our aligned results to be plugged into novel-view synthesis reconstruction methods to experience cartoons from viewpoints never drawn before. Our project page is https://toon3d.studio/.

Vector graphics are widely used in digital art and highly favored by designers due to their scalability and layer-wise properties. However, the process of creating and editing vector graphics requires creativity and design expertise, making it a time-consuming task. Recent advancements in text-to-vector (T2V) generation have aimed to make this process more accessible. However, existing T2V methods directly optimize control points of vector graphics paths, often resulting in intersecting or jagged paths due to the lack of geometry constraints. To overcome these limitations, we propose a novel neural path representation by designing a dual-branch Variational Autoencoder (VAE) that learns the path latent space from both sequence and image modalities. By optimizing the combination of neural paths, we can incorporate geometric constraints while preserving expressivity in generated SVGs. Furthermore, we introduce a two-stage path optimization method to improve the visual and topological quality of generated SVGs. In the first stage, a pre-trained text-to-image diffusion model guides the initial generation of complex vector graphics through the Variational Score Distillation (VSD) process. In the second stage, we refine the graphics using a layer-wise image vectorization strategy to achieve clearer elements and structure. We demonstrate the effectiveness of our method through extensive experiments and showcase various applications. The project page is https://intchous.github.io/T2V-NPR.

Visual In-Context Learning (ICL) has emerged as a promising research area due to its capability to accomplish various tasks with limited example pairs through analogical reasoning. However, training-based visual ICL has limitations in its ability to generalize to unseen tasks and requires the collection of a diverse task dataset. On the other hand, existing methods in the inference-based visual ICL category solely rely on textual prompts, which fail to capture fine-grained contextual information from given examples and can be time-consuming when converting from images to text prompts. To address these challenges, we propose Analogist, a novel inference-based visual ICL approach that exploits both visual and textual prompting techniques using a text-to-image diffusion model pretrained for image inpainting. For visual prompting, we propose a self-attention cloning (SAC) method to guide the fine-grained structural-level analogy between image examples. For textual prompting, we leverage GPT-4V's visual reasoning capability to efficiently generate text prompts and introduce a cross-attention masking (CAM) operation to enhance the accuracy of semantic-level analogy guided by text prompts. Our method is out-of-the-box and does not require fine-tuning or optimization. It is also generic and flexible, enabling a wide range of visual tasks to be performed in an in-context manner. Extensive experiments demonstrate the superiority of our method over existing approaches, both qualitatively and quantitatively.

Advances in 3D reconstruction have enabled high-quality 3D capture, but require a user to collect hundreds to thousands of images to create a 3D scene. We present CAT3D, a method for creating anything in 3D by simulating this real-world capture process with a multi-view diffusion model. Given any number of input images and a set of target novel viewpoints, our model generates highly consistent novel views of a scene. These generated views can be used as input to robust 3D reconstruction techniques to produce 3D representations that can be rendered from any viewpoint in real-time. CAT3D can create entire 3D scenes in as little as one minute, and outperforms existing methods for single image and few-view 3D scene creation. See our project page for results and interactive demos at https://cat3d.github.io .

We are living in a three-dimensional space while moving forward through a fourth dimension: time. To allow artificial intelligence to develop a comprehensive understanding of such a 4D environment, we introduce 4D Panoptic Scene Graph (PSG-4D), a new representation that bridges the raw visual data perceived in a dynamic 4D world and high-level visual understanding. Specifically, PSG-4D abstracts rich 4D sensory data into nodes, which represent entities with precise location and status information, and edges, which capture the temporal relations. To facilitate research in this new area, we build a richly annotated PSG-4D dataset consisting of 3K RGB-D videos with a total of 1M frames, each of which is labeled with 4D panoptic segmentation masks as well as fine-grained, dynamic scene graphs. To solve PSG-4D, we propose PSG4DFormer, a Transformer-based model that can predict panoptic segmentation masks, track masks along the time axis, and generate the corresponding scene graphs via a relation component. Extensive experiments on the new dataset show that our method can serve as a strong baseline for future research on PSG-4D. In the end, we provide a real-world application example to demonstrate how we can achieve dynamic scene understanding by integrating a large language model into our PSG-4D system.

This paper introduces Grounding DINO 1.5, a suite of advanced open-set object detection models developed by IDEA Research, which aims to advance the "Edge" of open-set object detection. The suite encompasses two models: Grounding DINO 1.5 Pro, a high-performance model designed for stronger generalization capability across a wide range of scenarios, and Grounding DINO 1.5 Edge, an efficient model optimized for faster speed demanded in many applications requiring edge deployment. The Grounding DINO 1.5 Pro model advances its predecessor by scaling up the model architecture, integrating an enhanced vision backbone, and expanding the training dataset to over 20 million images with grounding annotations, thereby achieving a richer semantic understanding. The Grounding DINO 1.5 Edge model, while designed for efficiency with reduced feature scales, maintains robust detection capabilities by being trained on the same comprehensive dataset. Empirical results demonstrate the effectiveness of Grounding DINO 1.5, with the Grounding DINO 1.5 Pro model attaining a 54.3 AP on the COCO detection benchmark and a 55.7 AP on the LVIS-minival zero-shot transfer benchmark, setting new records for open-set object detection. Furthermore, the Grounding DINO 1.5 Edge model, when optimized with TensorRT, achieves a speed of 75.2 FPS while attaining a zero-shot performance of 36.2 AP on the LVIS-minival benchmark, making it more suitable for edge computing scenarios. Model examples and demos with API will be released at https://github.com/IDEA-Research/Grounding-DINO-1.5-API

Large vision-language models (VLMs) fine-tuned on specialized visual instruction-following data have exhibited impressive language reasoning capabilities across various scenarios. However, this fine-tuning paradigm may not be able to efficiently learn optimal decision-making agents in multi-step goal-directed tasks from interactive environments. To address this challenge, we propose an algorithmic framework that fine-tunes VLMs with reinforcement learning (RL). Specifically, our framework provides a task description and then prompts the VLM to generate chain-of-thought (CoT) reasoning, enabling the VLM to efficiently explore intermediate reasoning steps that lead to the final text-based action. Next, the open-ended text output is parsed into an executable action to interact with the environment to obtain goal-directed task rewards. Finally, our framework uses these task rewards to fine-tune the entire VLM with RL. Empirically, we demonstrate that our proposed framework enhances the decision-making capabilities of VLM agents across various tasks, enabling 7b models to outperform commercial models such as GPT4-V or Gemini. Furthermore, we find that CoT reasoning is a crucial component for performance improvement, as removing the CoT reasoning results in a significant decrease in the overall performance of our method.

Despite noise and caption quality having been acknowledged as important factors impacting vision-language contrastive pre-training, in this paper, we show that the full potential of improving the training process by addressing such issues is yet to be realized. Specifically, we firstly study and analyze two issues affecting training: incorrect assignment of negative pairs, and low caption quality and diversity. Then, we devise effective solutions for addressing both problems, which essentially require training with multiple true positive pairs. Finally, we propose training with sigmoid loss to address such a requirement. We show very large gains over the current state-of-the-art for both image recognition ($\sim +6\%$ on average over 11 datasets) and image retrieval ($\sim +19\%$ on Flickr30k and $\sim +15\%$ on MSCOCO).

The goal of this work is to simultaneously generate natural talking faces and speech outputs from text. We achieve this by integrating Talking Face Generation (TFG) and Text-to-Speech (TTS) systems into a unified framework. We address the main challenges of each task: (1) generating a range of head poses representative of real-world scenarios, and (2) ensuring voice consistency despite variations in facial motion for the same identity. To tackle these issues, we introduce a motion sampler based on conditional flow matching, which is capable of high-quality motion code generation in an efficient way. Moreover, we introduce a novel conditioning method for the TTS system, which utilises motion-removed features from the TFG model to yield uniform speech outputs. Our extensive experiments demonstrate that our method effectively creates natural-looking talking faces and speech that accurately match the input text. To our knowledge, this is the first effort to build a multimodal synthesis system that can generalise to unseen identities.

In this work, our goals are two fold: large-vocabulary continuous sign language recognition (CSLR), and sign language retrieval. To this end, we introduce a multi-task Transformer model, CSLR2, that is able to ingest a signing sequence and output in a joint embedding space between signed language and spoken language text. To enable CSLR evaluation in the large-vocabulary setting, we introduce new dataset annotations that have been manually collected. These provide continuous sign-level annotations for six hours of test videos, and will be made publicly available. We demonstrate that by a careful choice of loss functions, training the model for both the CSLR and retrieval tasks is mutually beneficial in terms of performance -- retrieval improves CSLR performance by providing context, while CSLR improves retrieval with more fine-grained supervision. We further show the benefits of leveraging weak and noisy supervision from large-vocabulary datasets such as BOBSL, namely sign-level pseudo-labels, and English subtitles. Our model significantly outperforms the previous state of the art on both tasks.

This paper investigates the dynamics of a deep neural network (DNN) learning interactions. Previous studies have discovered and mathematically proven that given each input sample, a well-trained DNN usually only encodes a small number of interactions (non-linear relationships) between input variables in the sample. A series of theorems have been derived to prove that we can consider the DNN's inference equivalent to using these interactions as primitive patterns for inference. In this paper, we discover the DNN learns interactions in two phases. The first phase mainly penalizes interactions of medium and high orders, and the second phase mainly learns interactions of gradually increasing orders. We can consider the two-phase phenomenon as the starting point of a DNN learning over-fitted features. Such a phenomenon has been widely shared by DNNs with various architectures trained for different tasks. Therefore, the discovery of the two-phase dynamics provides a detailed mechanism for how a DNN gradually learns different inference patterns (interactions). In particular, we have also verified the claim that high-order interactions have weaker generalization power than low-order interactions. Thus, the discovered two-phase dynamics also explains how the generalization power of a DNN changes during the training process.

Deep learning models, particularly Convolutional Neural Networks (CNNs), have demonstrated exceptional performance in diagnosing skin diseases, often outperforming dermatologists. However, they have also unveiled biases linked to specific demographic traits, notably concerning diverse skin tones or gender, prompting concerns regarding fairness and limiting their widespread deployment. Researchers are actively working to ensure fairness in AI-based solutions, but existing methods incur an accuracy loss when striving for fairness. To solve this issue, we propose a `two-biased teachers' (i.e., biased on different sensitive attributes) based approach to transfer fair knowledge into the student network. Our approach mitigates biases present in the student network without harming its predictive accuracy. In fact, in most cases, our approach improves the accuracy of the baseline model. To achieve this goal, we developed a weighted loss function comprising biasing and debiasing loss terms. We surpassed available state-of-the-art approaches to attain fairness and also improved the accuracy at the same time. The proposed approach has been evaluated and validated on two dermatology datasets using standard accuracy and fairness evaluation measures. We will make source code publicly available to foster reproducibility and future research.

This paper introduces the Global-Local Image Perceptual Score (GLIPS), an image metric designed to assess the photorealistic image quality of AI-generated images with a high degree of alignment to human visual perception. Traditional metrics such as FID and KID scores do not align closely with human evaluations. The proposed metric incorporates advanced transformer-based attention mechanisms to assess local similarity and Maximum Mean Discrepancy (MMD) to evaluate global distributional similarity. To evaluate the performance of GLIPS, we conducted a human study on photorealistic image quality. Comprehensive tests across various generative models demonstrate that GLIPS consistently outperforms existing metrics like FID, SSIM, and MS-SSIM in terms of correlation with human scores. Additionally, we introduce the Interpolative Binning Scale (IBS), a refined scaling method that enhances the interpretability of metric scores by aligning them more closely with human evaluative standards. The proposed metric and scaling approach not only provides more reliable assessments of AI-generated images but also suggest pathways for future enhancements in image generation technologies.

As large language models (LLMs) evolve, their integration with 3D spatial data (3D-LLMs) has seen rapid progress, offering unprecedented capabilities for understanding and interacting with physical spaces. This survey provides a comprehensive overview of the methodologies enabling LLMs to process, understand, and generate 3D data. Highlighting the unique advantages of LLMs, such as in-context learning, step-by-step reasoning, open-vocabulary capabilities, and extensive world knowledge, we underscore their potential to significantly advance spatial comprehension and interaction within embodied Artificial Intelligence (AI) systems. Our investigation spans various 3D data representations, from point clouds to Neural Radiance Fields (NeRFs). It examines their integration with LLMs for tasks such as 3D scene understanding, captioning, question-answering, and dialogue, as well as LLM-based agents for spatial reasoning, planning, and navigation. The paper also includes a brief review of other methods that integrate 3D and language. The meta-analysis presented in this paper reveals significant progress yet underscores the necessity for novel approaches to harness the full potential of 3D-LLMs. Hence, with this paper, we aim to chart a course for future research that explores and expands the capabilities of 3D-LLMs in understanding and interacting with the complex 3D world. To support this survey, we have established a project page where papers related to our topic are organized and listed: https://github.com/ActiveVisionLab/Awesome-LLM-3D.

Foundation models in computational pathology promise to unlock the development of new clinical decision support systems and models for precision medicine. However, there is a mismatch between most clinical analysis, which is defined at the level of one or more whole slide images, and foundation models to date, which process the thousands of image tiles contained in a whole slide image separately. The requirement to train a network to aggregate information across a large number of tiles in multiple whole slide images limits these models' impact. In this work, we present a slide-level foundation model for H&E-stained histopathology, PRISM, that builds on Virchow tile embeddings and leverages clinical report text for pre-training. Using the tile embeddings, PRISM produces slide-level embeddings with the ability to generate clinical reports, resulting in several modes of use. Using text prompts, PRISM achieves zero-shot cancer detection and sub-typing performance approaching and surpassing that of a supervised aggregator model. Using the slide embeddings with linear classifiers, PRISM surpasses supervised aggregator models. Furthermore, we demonstrate that fine-tuning of the PRISM slide encoder yields label-efficient training for biomarker prediction, a task that typically suffers from low availability of training data; an aggregator initialized with PRISM and trained on as little as 10% of the training data can outperform a supervised baseline that uses all of the data.

Esophageal cancer is one of the most common types of cancer worldwide and ranks sixth in cancer-related mortality. Accurate computer-assisted diagnosis of cancer progression can help physicians effectively customize personalized treatment plans. Currently, CT-based cancer diagnosis methods have received much attention for their comprehensive ability to examine patients' conditions. However, multi-modal based methods may likely introduce information redundancy, leading to underperformance. In addition, efficient and effective interactions between multi-modal representations need to be further explored, lacking insightful exploration of prognostic correlation in multi-modality features. In this work, we introduce a multi-modal heterogeneous graph-based conditional feature-guided diffusion model for lymph node metastasis diagnosis based on CT images as well as clinical measurements and radiomics data. To explore the intricate relationships between multi-modal features, we construct a heterogeneous graph. Following this, a conditional feature-guided diffusion approach is applied to eliminate information redundancy. Moreover, we propose a masked relational representation learning strategy, aiming to uncover the latent prognostic correlations and priorities of primary tumor and lymph node image representations. Various experimental results validate the effectiveness of our proposed method. The code is available at https://github.com/wuchengyu123/MMFusion.

Brain lesion segmentation plays an essential role in neurological research and diagnosis. As brain lesions can be caused by various pathological alterations, different types of brain lesions tend to manifest with different characteristics on different imaging modalities. Due to this complexity, brain lesion segmentation methods are often developed in a task-specific manner. A specific segmentation model is developed for a particular lesion type and imaging modality. However, the use of task-specific models requires predetermination of the lesion type and imaging modality, which complicates their deployment in real-world scenarios. In this work, we propose a universal foundation model for 3D brain lesion segmentation, which can automatically segment different types of brain lesions for input data of various imaging modalities. We formulate a novel Mixture of Modality Experts (MoME) framework with multiple expert networks attending to different imaging modalities. A hierarchical gating network combines the expert predictions and fosters expertise collaboration. Furthermore, we introduce a curriculum learning strategy during training to avoid the degeneration of each expert network and preserve their specialization. We evaluated the proposed method on nine brain lesion datasets, encompassing five imaging modalities and eight lesion types. The results show that our model outperforms state-of-the-art universal models and provides promising generalization to unseen datasets.

We identify an issue in multi-task learnable compression, in which a representation learned for one task does not positively contribute to the rate-distortion performance of a different task as much as expected, given the estimated amount of information available in it. We interpret this issue using the predictive $\mathcal{V}$-information framework. In learnable scalable coding, previous work increased the utilization of side-information for input reconstruction by also rewarding input reconstruction when learning this shared representation. We evaluate the impact of this idea in the context of input reconstruction more rigorously and extended it to other computer vision tasks. We perform experiments using representations trained for object detection on COCO 2017 and depth estimation on the Cityscapes dataset, and use them to assist in image reconstruction and semantic segmentation tasks. The results show considerable improvements in the rate-distortion performance of the assisted tasks. Moreover, using the proposed representations, the performance of the base tasks are also improved. Results suggest that the proposed method induces simpler representations that are more compatible with downstream processes.

Instance segmentation is data-hungry, and as model capacity increases, data scale becomes crucial for improving the accuracy. Most instance segmentation datasets today require costly manual annotation, limiting their data scale. Models trained on such data are prone to overfitting on the training set, especially for those rare categories. While recent works have delved into exploiting generative models to create synthetic datasets for data augmentation, these approaches do not efficiently harness the full potential of generative models. To address these issues, we introduce a more efficient strategy to construct generative datasets for data augmentation, termed DiverGen. Firstly, we provide an explanation of the role of generative data from the perspective of distribution discrepancy. We investigate the impact of different data on the distribution learned by the model. We argue that generative data can expand the data distribution that the model can learn, thus mitigating overfitting. Additionally, we find that the diversity of generative data is crucial for improving model performance and enhance it through various strategies, including category diversity, prompt diversity, and generative model diversity. With these strategies, we can scale the data to millions while maintaining the trend of model performance improvement. On the LVIS dataset, DiverGen significantly outperforms the strong model X-Paste, achieving +1.1 box AP and +1.1 mask AP across all categories, and +1.9 box AP and +2.5 mask AP for rare categories.

Point cloud segmentation (PCS) plays an essential role in robot perception and navigation tasks. To efficiently understand large-scale outdoor point clouds, their range image representation is commonly adopted. This image-like representation is compact and structured, making range image-based PCS models practical. However, undesirable missing values in the range images damage the shapes and patterns of objects. This problem creates difficulty for the models in learning coherent and complete geometric information from the objects. Consequently, the PCS models only achieve inferior performance. Delving deeply into this issue, we find that the use of unreasonable projection approaches and deskewing scans mainly leads to unwanted missing values in the range images. Besides, almost all previous works fail to consider filling in the unexpected missing values in the PCS task. To alleviate this problem, we first propose a new projection method, namely scan unfolding++ (SU++), to avoid massive missing values in the generated range images. Then, we introduce a simple yet effective approach, namely range-dependent $K$-nearest neighbor interpolation ($K$NNI), to further fill in missing values. Finally, we introduce the Filling Missing Values Network (FMVNet) and Fast FMVNet. Extensive experimental results on SemanticKITTI, SemanticPOSS, and nuScenes datasets demonstrate that by employing the proposed SU++ and $K$NNI, existing range image-based PCS models consistently achieve better performance than the baseline models. Besides, both FMVNet and Fast FMVNet achieve state-of-the-art performance in terms of the speed-accuracy trade-off. The proposed methods can be applied to other range image-based tasks and practical applications.

Remote sensing image-text retrieval constitutes a foundational aspect of remote sensing interpretation tasks, facilitating the alignment of vision and language representations. This paper introduces a prior instruction representation (PIR) learning paradigm that draws on prior knowledge to instruct adaptive learning of vision and text representations. Based on PIR, a domain-adapted remote sensing image-text retrieval framework PIR-ITR is designed to address semantic noise issues in vision-language understanding tasks. However, with massive additional data for pre-training the vision-language foundation model, remote sensing image-text retrieval is further developed into an open-domain retrieval task. Continuing with the above, we propose PIR-CLIP, a domain-specific CLIP-based framework for remote sensing image-text retrieval, to address semantic noise in remote sensing vision-language representations and further improve open-domain retrieval performance. In vision representation, Vision Instruction Representation (VIR) based on Spatial-PAE utilizes the prior-guided knowledge of the remote sensing scene recognition by building a belief matrix to select key features for reducing the impact of semantic noise. In text representation, Language Cycle Attention (LCA) based on Temporal-PAE uses the previous time step to cyclically activate the current time step to enhance text representation capability. A cluster-wise Affiliation Loss (AL) is proposed to constrain the inter-classes and to reduce the semantic confusion zones in the common subspace. Comprehensive experiments demonstrate that PIR could enhance vision and text representations and outperform the state-of-the-art methods of closed-domain and open-domain retrieval on two benchmark datasets, RSICD and RSITMD.

Hyperspectral target detection (HTD) aims to identify specific materials based on spectral information in hyperspectral imagery and can detect point targets, some of which occupy a smaller than one-pixel area. However, existing HTD methods are developed based on per-pixel binary classification, which limits the feature representation capability for point targets. In this paper, we rethink the hyperspectral point target detection from the object detection perspective, and focus more on the object-level prediction capability rather than the pixel classification capability. Inspired by the token-based processing flow of Detection Transformer (DETR), we propose the first specialized network for hyperspectral multi-class point object detection, SpecDETR. Without the backbone part of the current object detection framework, SpecDETR treats the spectral features of each pixel in hyperspectral images as a token and utilizes a multi-layer Transformer encoder with local and global coordination attention modules to extract deep spatial-spectral joint features. SpecDETR regards point object detection as a one-to-many set prediction problem, thereby achieving a concise and efficient DETR decoder that surpasses the current state-of-the-art DETR decoder in terms of parameters and accuracy in point object detection. We develop a simulated hyperSpectral Point Object Detection benchmark termed SPOD, and for the first time, evaluate and compare the performance of current object detection networks and HTD methods on hyperspectral multi-class point object detection. SpecDETR demonstrates superior performance as compared to current object detection networks and HTD methods on the SPOD dataset. Additionally, we validate on a public HTD dataset that by using data simulation instead of manual annotation, SpecDETR can detect real-world single-spectral point objects directly.

In this work, we introduce Libra, a prototype model with a decoupled vision system on a large language model (LLM). The decoupled vision system decouples inner-modal modeling and cross-modal interaction, yielding unique visual information modeling and effective cross-modal comprehension. Libra is trained through discrete auto-regressive modeling on both vision and language inputs. Specifically, we incorporate a routed visual expert with a cross-modal bridge module into a pretrained LLM to route the vision and language flows during attention computing to enable different attention patterns in inner-modal modeling and cross-modal interaction scenarios. Experimental results demonstrate that the dedicated design of Libra achieves a strong MLLM baseline that rivals existing works in the image-to-text scenario with merely 50 million training data, providing a new perspective for future multimodal foundation models. Code is available at https://github.com/YifanXu74/Libra.

In the typical urban intersection scenario, both vehicles and infrastructures are equipped with visual and LiDAR sensors. By successfully integrating the data from vehicle-side and road monitoring devices, a more comprehensive and accurate environmental perception and information acquisition can be achieved. The Calibration of sensors, as an essential component of autonomous driving technology, has consistently drawn significant attention. Particularly in scenarios involving multiple sensors collaboratively perceiving and addressing localization challenges, the requirement for inter-sensor calibration becomes crucial. Recent years have witnessed the emergence of the concept of multi-end cooperation, where infrastructure captures and transmits surrounding environment information to vehicles, bolstering their perception capabilities while mitigating costs. However, this also poses technical complexities, underscoring the pressing need for diverse end calibration. Camera and LiDAR, the bedrock sensors in autonomous driving, exhibit expansive applicability. This paper comprehensively examines and analyzes the calibration of multi-end camera-LiDAR setups from vehicle, roadside, and vehicle-road cooperation perspectives, outlining their relevant applications and profound significance. Concluding with a summary, we present our future-oriented ideas and hypotheses.

Multistep instructions, such as recipes and how-to guides, greatly benefit from visual aids, such as a series of images that accompany the instruction steps. While Large Language Models (LLMs) have become adept at generating coherent textual steps, Large Vision/Language Models (LVLMs) are less capable of generating accompanying image sequences. The most challenging aspect is that each generated image needs to adhere to the relevant textual step instruction, as well as be visually consistent with earlier images in the sequence. To address this problem, we propose an approach for generating consistent image sequences, which integrates a Latent Diffusion Model (LDM) with an LLM to transform the sequence into a caption to maintain the semantic coherence of the sequence. In addition, to maintain the visual coherence of the image sequence, we introduce a copy mechanism to initialise reverse diffusion processes with a latent vector iteration from a previously generated image from a relevant step. Both strategies will condition the reverse diffusion process on the sequence of instruction steps and tie the contents of the current image to previous instruction steps and corresponding images. Experiments show that the proposed approach is preferred by humans in 46.6% of the cases against 26.6% for the second best method. In addition, automatic metrics showed that the proposed method maintains semantic coherence and visual consistency across steps in both domains.

In this paper, we propose an integrated framework for multi-granular explanation of video summarization. This framework integrates methods for producing explanations both at the fragment level (indicating which video fragments influenced the most the decisions of the summarizer) and the more fine-grained visual object level (highlighting which visual objects were the most influential for the summarizer). To build this framework, we extend our previous work on this field, by investigating the use of a model-agnostic, perturbation-based approach for fragment-level explanation of the video summarization results, and introducing a new method that combines the results of video panoptic segmentation with an adaptation of a perturbation-based explanation approach to produce object-level explanations. The performance of the developed framework is evaluated using a state-of-the-art summarization method and two datasets for benchmarking video summarization. The findings of the conducted quantitative and qualitative evaluations demonstrate the ability of our framework to spot the most and least influential fragments and visual objects of the video for the summarizer, and to provide a comprehensive set of visual-based explanations about the output of the summarization process.

Natural language could play an important role in developing generalist surgical models by providing a broad source of supervision from raw texts. This flexible form of supervision can enable the model's transferability across datasets and tasks as natural language can be used to reference learned visual concepts or describe new ones. In this work, we present HecVL, a novel hierarchical video-language pretraining approach for building a generalist surgical model. Specifically, we construct a hierarchical video-text paired dataset by pairing the surgical lecture video with three hierarchical levels of texts: at clip-level, atomic actions using transcribed audio texts; at phase-level, conceptual text summaries; and at video-level, overall abstract text of the surgical procedure. Then, we propose a novel fine-to-coarse contrastive learning framework that learns separate embedding spaces for the three video-text hierarchies using a single model. By disentangling embedding spaces of different hierarchical levels, the learned multi-modal representations encode short-term and long-term surgical concepts in the same model. Thanks to the injected textual semantics, we demonstrate that the HecVL approach can enable zero-shot surgical phase recognition without any human annotation. Furthermore, we show that the same HecVL model for surgical phase recognition can be transferred across different surgical procedures and medical centers.

Deformable image registration (alignment) is highly sought after in numerous clinical applications, such as computer aided diagnosis and disease progression analysis. Deep Convolutional Neural Network (DCNN)-based image registration methods have demonstrated advantages in terms of registration accuracy and computational speed. However, while most methods excel at global alignment, they often perform worse in aligning local regions. To address this challenge, this paper proposes a mask-guided encoder-decoder DCNN-based image registration method, named as MrRegNet. This approach employs a multi-resolution encoder for feature extraction and subsequently estimates multi-resolution displacement fields in the decoder to handle the substantial deformation of images. Furthermore, segmentation masks are employed to direct the model's attention toward aligning local regions. The results show that the proposed method outperforms traditional methods like Demons and a well-known deep learning method, VoxelMorph, on a public 3D brain MRI dataset (OASIS) and a local 2D brain MRI dataset with large deformations. Importantly, the image alignment accuracies are significantly improved at local regions guided by segmentation masks. Github link:https://github.com/ruizhe-l/MrRegNet.

Open-vocabulary object detection (OvOD) has transformed detection into a language-guided task, empowering users to freely define their class vocabularies of interest during inference. However, our initial investigation indicates that existing OvOD detectors exhibit significant variability when dealing with vocabularies across various semantic granularities, posing a concern for real-world deployment. To this end, we introduce Semantic Hierarchy Nexus (SHiNe), a novel classifier that uses semantic knowledge from class hierarchies. It runs offline in three steps: i) it retrieves relevant super-/sub-categories from a hierarchy for each target class; ii) it integrates these categories into hierarchy-aware sentences; iii) it fuses these sentence embeddings to generate the nexus classifier vector. Our evaluation on various detection benchmarks demonstrates that SHiNe enhances robustness across diverse vocabulary granularities, achieving up to +31.9% mAP50 with ground truth hierarchies, while retaining improvements using hierarchies generated by large language models. Moreover, when applied to open-vocabulary classification on ImageNet-1k, SHiNe improves the CLIP zero-shot baseline by +2.8% accuracy. SHiNe is training-free and can be seamlessly integrated with any off-the-shelf OvOD detector, without incurring additional computational overhead during inference. The code is open source.

In recent years considerable research in LiDAR semantic segmentation was conducted, introducing several new state of the art models. However, most research focuses on single-scan point clouds, limiting performance especially in long distance outdoor scenarios, by omitting time-sequential information. Moreover, varying-density and occlusions constitute significant challenges in single-scan approaches. In this paper we propose a LiDAR point cloud preprocessing and postprocessing method. This multi-stage approach, in conjunction with state of the art models in a multi-scan setting, aims to solve those challenges. We demonstrate the benefits of our method through quantitative evaluation with the given models in single-scan settings. In particular, we achieve significant improvements in mIoU performance of over 5 percentage point in medium range and over 10 percentage point in far range. This is essential for 3D semantic scene understanding in long distance as well as for applications where offline processing is permissible.

Hierarchical leaf vein segmentation is a crucial but under-explored task in agricultural sciences, where analysis of the hierarchical structure of plant leaf venation can contribute to plant breeding. While current segmentation techniques rely on data-driven models, there is no publicly available dataset specifically designed for hierarchical leaf vein segmentation. To address this gap, we introduce the HierArchical Leaf Vein Segmentation (HALVS) dataset, the first public hierarchical leaf vein segmentation dataset. HALVS comprises 5,057 real-scanned high-resolution leaf images collected from three plant species: soybean, sweet cherry, and London planetree. It also includes human-annotated ground truth for three orders of leaf veins, with a total labeling effort of 83.8 person-days. Based on HALVS, we further develop a label-efficient learning paradigm that leverages partial label information, i.e. missing annotations for tertiary veins. Empirical studies are performed on HALVS, revealing new observations, challenges, and research directions on leaf vein segmentation.

Event Stream Super-Resolution (ESR) aims to address the challenge of insufficient spatial resolution in event streams, which holds great significance for the application of event cameras in complex scenarios. Previous works for ESR often process positive and negative events in a mixed paradigm. This paradigm limits their ability to effectively model the unique characteristics of each event and mutually refine each other by considering their correlations. In this paper, we propose a bilateral event mining and complementary network (BMCNet) to fully leverage the potential of each event and capture the shared information to complement each other simultaneously. Specifically, we resort to a two-stream network to accomplish comprehensive mining of each type of events individually. To facilitate the exchange of information between two streams, we propose a bilateral information exchange (BIE) module. This module is layer-wisely embedded between two streams, enabling the effective propagation of hierarchical global information while alleviating the impact of invalid information brought by inherent characteristics of events. The experimental results demonstrate that our approach outperforms the previous state-of-the-art methods in ESR, achieving performance improvements of over 11\% on both real and synthetic datasets. Moreover, our method significantly enhances the performance of event-based downstream tasks such as object recognition and video reconstruction. Our code is available at https://github.com/Lqm26/BMCNet-ESR.

Remote sensing image dehazing (RSID) aims to remove nonuniform and physically irregular haze factors for high-quality image restoration. The emergence of CNNs and Transformers has taken extraordinary strides in the RSID arena. However, these methods often struggle to demonstrate the balance of adequate long-range dependency modeling and maintaining computational efficiency. To this end, we propose the first lightweight network on the mamba-based model called RSDhamba in the field of RSID. Greatly inspired by the recent rise of Selective State Space Model (SSM) for its superior performance in modeling linear complexity and remote dependencies, our designed RSDehamba integrates the SSM framework into the U-Net architecture. Specifically, we propose the Vision Dehamba Block (VDB) as the core component of the overall network, which utilizes the linear complexity of SSM to achieve the capability of global context encoding. Simultaneously, the Direction-aware Scan Module (DSM) is designed to dynamically aggregate feature exchanges over different directional domains to effectively enhance the flexibility of sensing the spatially varying distribution of haze. In this way, our RSDhamba fully demonstrates the superiority of spatial distance capture dependencies and channel information exchange for better extraction of haze features. Extensive experimental results on widely used benchmarks validate the surpassing performance of our RSDehamba against existing state-of-the-art methods.

A recent study by De et al. (2022) has reported that large-scale representation learning through pre-training on a public dataset significantly enhances differentially private (DP) learning in downstream tasks, despite the high dimensionality of the feature space. To theoretically explain this phenomenon, we consider the setting of a layer-peeled model in representation learning, which results in interesting phenomena related to learned features in deep learning and transfer learning, known as Neural Collapse (NC). Within the framework of NC, we establish an error bound indicating that the misclassification error is independent of dimension when the distance between actual features and the ideal ones is smaller than a threshold. Additionally, the quality of the features in the last layer is empirically evaluated under different pre-trained models within the framework of NC, showing that a more powerful transformer leads to a better feature representation. Furthermore, we reveal that DP fine-tuning is less robust compared to fine-tuning without DP, particularly in the presence of perturbations. These observations are supported by both theoretical analyses and experimental evaluation. Moreover, to enhance the robustness of DP fine-tuning, we suggest several strategies, such as feature normalization or employing dimension reduction methods like Principal Component Analysis (PCA). Empirically, we demonstrate a significant improvement in testing accuracy by conducting PCA on the last-layer features.

The main challenge in learning image-conditioned robotic policies is acquiring a visual representation conducive to low-level control. Due to the high dimensionality of the image space, learning a good visual representation requires a considerable amount of visual data. However, when learning in the real world, data is expensive. Sim2Real is a promising paradigm for overcoming data scarcity in the real-world target domain by using a simulator to collect large amounts of cheap data closely related to the target task. However, it is difficult to transfer an image-conditioned policy from sim to real when the domains are very visually dissimilar. To bridge the sim2real visual gap, we propose using natural language descriptions of images as a unifying signal across domains that captures the underlying task-relevant semantics. Our key insight is that if two image observations from different domains are labeled with similar language, the policy should predict similar action distributions for both images. We demonstrate that training the image encoder to predict the language description or the distance between descriptions of a sim or real image serves as a useful, data-efficient pretraining step that helps learn a domain-invariant image representation. We can then use this image encoder as the backbone of an IL policy trained simultaneously on a large amount of simulated and a handful of real demonstrations. Our approach outperforms widely used prior sim2real methods and strong vision-language pretraining baselines like CLIP and R3M by 25 to 40%.

The performance of single image super-resolution depends heavily on how to generate and complement high-frequency details to low-resolution images. Recently, diffusion-based models exhibit great potential in generating high-quality images for super-resolution tasks. However, existing models encounter difficulties in directly predicting high-frequency information of wide bandwidth by solely utilizing the high-resolution ground truth as the target for all sampling timesteps. To tackle this problem and achieve higher-quality super-resolution, we propose a novel Frequency Domain-guided multiscale Diffusion model (FDDiff), which decomposes the high-frequency information complementing process into finer-grained steps. In particular, a wavelet packet-based frequency complement chain is developed to provide multiscale intermediate targets with increasing bandwidth for reverse diffusion process. Then FDDiff guides reverse diffusion process to progressively complement the missing high-frequency details over timesteps. Moreover, we design a multiscale frequency refinement network to predict the required high-frequency components at multiple scales within one unified network. Comprehensive evaluations on popular benchmarks are conducted, and demonstrate that FDDiff outperforms prior generative methods with higher-fidelity super-resolution results.

The accelerated progress of artificial intelligence (AI) has popularized deep learning models across domains, yet their inherent opacity poses challenges, notably in critical fields like healthcare, medicine and the geosciences. Explainable AI (XAI) has emerged to shed light on these "black box" models, helping decipher their decision making process. Nevertheless, different XAI methods yield highly different explanations. This inter-method variability increases uncertainty and lowers trust in deep networks' predictions. In this study, for the first time, we propose a novel framework designed to enhance the explainability of deep networks, by maximizing both the accuracy and the comprehensibility of the explanations. Our framework integrates various explanations from established XAI methods and employs a non-linear "explanation optimizer" to construct a unique and optimal explanation. Through experiments on multi-class and binary classification tasks in 2D object and 3D neuroscience imaging, we validate the efficacy of our approach. Our explanation optimizer achieved superior faithfulness scores, averaging 155% and 63% higher than the best performing XAI method in the 3D and 2D applications, respectively. Additionally, our approach yielded lower complexity, increasing comprehensibility. Our results suggest that optimal explanations based on specific criteria are derivable and address the issue of inter-method variability in the current XAI literature.

Automated medical image analysis systems often require large amounts of training data with high quality labels, which are difficult and time consuming to generate. This paper introduces Radiology Object in COntext version 2 (ROCOv2), a multimodal dataset consisting of radiological images and associated medical concepts and captions extracted from the PMC Open Access subset. It is an updated version of the ROCO dataset published in 2018, and adds 35,705 new images added to PMC since 2018. It further provides manually curated concepts for imaging modalities with additional anatomical and directional concepts for X-rays. The dataset consists of 79,789 images and has been used, with minor modifications, in the concept detection and caption prediction tasks of ImageCLEFmedical Caption 2023. The dataset is suitable for training image annotation models based on image-caption pairs, or for multi-label image classification using Unified Medical Language System (UMLS) concepts provided with each image. In addition, it can serve for pre-training of medical domain models, and evaluation of deep learning models for multi-task learning.

Real driving-video dehazing poses a significant challenge due to the inherent difficulty in acquiring precisely aligned hazy/clear video pairs for effective model training, especially in dynamic driving scenarios with unpredictable weather conditions. In this paper, we propose a pioneering approach that addresses this challenge through a nonaligned regularization strategy. Our core concept involves identifying clear frames that closely match hazy frames, serving as references to supervise a video dehazing network. Our approach comprises two key components: reference matching and video dehazing. Firstly, we introduce a non-aligned reference frame matching module, leveraging an adaptive sliding window to match high-quality reference frames from clear videos. Video dehazing incorporates flow-guided cosine attention sampler and deformable cosine attention fusion modules to enhance spatial multiframe alignment and fuse their improved information. To validate our approach, we collect a GoProHazy dataset captured effortlessly with GoPro cameras in diverse rural and urban road environments. Extensive experiments demonstrate the superiority of the proposed method over current state-of-the-art methods in the challenging task of real driving-video dehazing. Project page.

Large pretrained transformers are increasingly being developed as generalised foundation models which can underpin powerful task-specific artificial intelligence models. Histopathology foundation models show promise across many tasks, but analyses have been limited by arbitrary hyperparameters that were not tuned to the specific task/dataset. We report the most rigorous single-task validation conducted to date of a histopathology foundation model, and the first performed in ovarian cancer subtyping. Attention-based multiple instance learning classifiers were compared using vision transformer and ResNet features generated through varied preprocessing and pretraining procedures. The training set consisted of 1864 whole slide images from 434 ovarian carcinoma cases at Leeds Hospitals. Five-class classification performance was evaluated through five-fold cross-validation, and these cross-validation models were ensembled for evaluation on a hold-out test set and an external set from the Transcanadian study. Reporting followed the TRIPOD+AI checklist. The vision transformer-based histopathology foundation model, UNI, performed best in every evaluation, with five-class balanced accuracies of 88% and 93% in hold-out internal and external testing, compared to the best ResNet model scores of 68% and 81%, respectively. Normalisations and augmentations aided the generalisability of ResNet-based models, but these still did not match the performance of UNI, which gave the best external performance in any ovarian cancer subtyping study to date. Histopathology foundation models offer a clear benefit to subtyping, improving classification performance to a degree where clinical utility is tangible, albeit with an increased computational burden. Such models could provide a second opinion in challenging cases and may improve the accuracy, objectivity, and efficiency of pathological diagnoses overall.

Due to the significant advances in large-scale text-to-image generation by diffusion model (DM), controllable human image generation has been attracting much attention recently. Existing works, such as Controlnet [36], T2I-adapter [20] and HumanSD [10] have demonstrated good abilities in generating human images based on pose conditions, they still fail to meet the requirements of real e-commerce scenarios. These include (1) the interaction between the shown product and human should be considered, (2) human parts like face/hand/arm/foot and the interaction between human model and product should be hyper-realistic, and (3) the identity of the product shown in advertising should be exactly consistent with the product itself. To this end, in this paper, we first define a new human image generation task for e-commerce marketing, i.e., Object-ID-retentive Human-object Interaction image Generation (OHG), and then propose a VirtualModel framework to generate human images for product shown, which supports displays of any categories of products and any types of human-object interaction. As shown in Figure 1, VirtualModel not only outperforms other methods in terms of accurate pose control and image quality but also allows for the display of user-specified product objects by maintaining the product-ID consistency and enhancing the plausibility of human-object interaction. Codes and data will be released.

Multi-modal Large Language Models (MLLMs) have recently achieved enhanced performance across various vision-language tasks including visual grounding capabilities. However, the adversarial robustness of visual grounding remains unexplored in MLLMs. To fill this gap, we use referring expression comprehension (REC) as an example task in visual grounding and propose three adversarial attack paradigms as follows. Firstly, untargeted adversarial attacks induce MLLMs to generate incorrect bounding boxes for each object. Besides, exclusive targeted adversarial attacks cause all generated outputs to the same target bounding box. In addition, permuted targeted adversarial attacks aim to permute all bounding boxes among different objects within a single image. Extensive experiments demonstrate that the proposed methods can successfully attack visual grounding capabilities of MLLMs. Our methods not only provide a new perspective for designing novel attacks but also serve as a strong baseline for improving the adversarial robustness for visual grounding of MLLMs.

In the new paradigm of semantic communication (SC), the focus is on delivering meanings behind bits by extracting semantic information from raw data. Recent advances in data-to-text models facilitate language-oriented SC, particularly for text-transformed image communication via image-to-text (I2T) encoding and text-to-image (T2I) decoding. However, although semantically aligned, the text is too coarse to precisely capture sophisticated visual features such as spatial locations, color, and texture, incurring a significant perceptual difference between intended and reconstructed images. To address this limitation, in this paper, we propose a novel language-oriented SC framework that communicates both text and a compressed image embedding and combines them using a latent diffusion model to reconstruct the intended image. Experimental results validate the potential of our approach, which transmits only 2.09\% of the original image size while achieving higher perceptual similarities in noisy communication channels compared to a baseline SC method that communicates only through text.The code is available at https://github.com/ispamm/Img2Img-SC/ .

With the development of deep neural network generative models in recent years, significant progress has been made in the research of depth estimation in lane scenes. However, current research achievements are mainly focused on clear daytime scenarios. In complex rainy environments, the influence of rain streaks and local fog effects often leads to erroneous increases in the overall depth estimation values in images. Moreover, these natural factors can introduce disturbances to the accurate prediction of depth boundaries in images. In this paper, we investigate lane depth estimation in complex rainy environments. Based on the concept of convolutional kernel prediction, we propose a dual-layer pixel-wise convolutional kernel prediction network trained on offline data. By predicting two sets of independent convolutional kernels for the target image, we restore the depth information loss caused by complex environmental factors and address the issue of rain streak artifacts generated by a single convolutional kernel set. Furthermore, considering the lack of real rainy lane data currently available, we introduce an image synthesis algorithm, RCFLane, which comprehensively considers the darkening of the environment due to rainfall and local fog effects. We create a synthetic dataset containing 820 experimental images, which we refer to as RainKITTI, on the commonly used depth estimation dataset KITTI. Extensive experiments demonstrate that our proposed depth estimation framework achieves favorable results in highly complex lane rainy environments.

Intraoperative ultrasound (iUS) imaging has the potential to improve surgical outcomes in brain surgery. However, its interpretation is challenging, even for expert neurosurgeons. In this work, we designed the first patient-specific framework that performs brain tumor segmentation in trackerless iUS. To disambiguate ultrasound imaging and adapt to the neurosurgeon's surgical objective, a patient-specific real-time network is trained using synthetic ultrasound data generated by simulating virtual iUS sweep acquisitions in pre-operative MR data. Extensive experiments performed in real ultrasound data demonstrate the effectiveness of the proposed approach, allowing for adapting to the surgeon's definition of surgical targets and outperforming non-patient-specific models, neurosurgeon experts, and high-end tracking systems. Our code is available at: \url{https://github.com/ReubenDo/MHVAE-Seg}.

The maturity classification of specialty crops such as strawberries and tomatoes is an essential agricultural downstream activity for selective harvesting and quality control (QC) at production and packaging sites. Recent advancements in Deep Learning (DL) have produced encouraging results in color images for maturity classification applications. However, hyperspectral imaging (HSI) outperforms methods based on color vision. Multivariate analysis methods and Convolutional Neural Networks (CNN) deliver promising results; however, a large amount of input data and the associated preprocessing requirements cause hindrances in practical application. Conventionally, the reflectance intensity in a given electromagnetic spectrum is employed in estimating fruit maturity. We present a feature extraction method to empirically demonstrate that the peak reflectance in subbands such as 500-670 nm (pigment band) and the wavelength of the peak position, and contrarily, the trough reflectance and its corresponding wavelength within 671-790 nm (chlorophyll band) are convenient to compute yet distinctive features for the maturity classification. The proposed feature selection method is beneficial because preprocessing, such as dimensionality reduction, is avoided before every prediction. The feature set is designed to capture these traits. The best SOTA methods, among 3D-CNN, 1D-CNN, and SVM, achieve at most 90.0 % accuracy for strawberries and 92.0 % for tomatoes on our dataset. Results show that the proposed method outperforms the SOTA as it yields an accuracy above 98.0 % in strawberry and 96.0 % in tomato classification. A comparative analysis of the time efficiency of these methods is also conducted, which shows the proposed method performs prediction at 13 Frames Per Second (FPS) compared to the maximum 1.16 FPS attained by the full-spectrum SVM classifier.

Bounding box regression is one of the important steps of object detection. However, rotation detectors often involve a more complicated loss based on SkewIoU which is unfriendly to gradient-based training. Most of the existing loss functions for rotated object detection calculate the difference between two bounding boxes only focus on the deviation of area or each points distance (e.g., $\mathcal{L}_{Smooth-\ell 1}$, $\mathcal{L}_{RotatedIoU}$ and $\mathcal{L}_{PIoU}$). The calculation process of some loss functions is extremely complex (e.g. $\mathcal{L}_{KFIoU}$). In order to improve the efficiency and accuracy of bounding box regression for rotated object detection, we proposed a novel metric for arbitrary shapes comparison based on minimum points distance, which takes most of the factors from existing loss functions for rotated object detection into account, i.e., the overlap or nonoverlapping area, the central points distance and the rotation angle. We also proposed a loss function called $\mathcal{L}_{FPDIoU}$ based on four points distance for accurate bounding box regression focusing on faster and high quality anchor boxes. In the experiments, $FPDIoU$ loss has been applied to state-of-the-art rotated object detection (e.g., RTMDET, H2RBox) models training with three popular benchmarks of rotated object detection including DOTA, DIOR, HRSC2016 and two benchmarks of arbitrary orientation scene text detection including ICDAR 2017 RRC-MLT and ICDAR 2019 RRC-MLT, which achieves better performance than existing loss functions.

Multiple-instance learning (MIL) is an attractive approach for digital pathology applications as it reduces the costs related to data collection and labelling. However, it is not clear how sensitive MIL is to clinically realistic domain shifts, i.e., differences in data distribution that could negatively affect performance, and if already existing metrics for detecting domain shifts work well with these algorithms. We trained an attention-based MIL algorithm to classify whether a whole-slide image of a lymph node contains breast tumour metastases. The algorithm was evaluated on data from a hospital in a different country and various subsets of this data that correspond to different levels of domain shift. Our contributions include showing that MIL for digital pathology is affected by clinically realistic differences in data, evaluating which features from a MIL model are most suitable for detecting changes in performance, and proposing an unsupervised metric named Fr\'echet Domain Distance (FDD) for quantification of domain shifts. Shift measure performance was evaluated through the mean Pearson correlation to change in classification performance, where FDD achieved 0.70 on 10-fold cross-validation models. The baselines included Deep ensemble, Difference of Confidence, and Representation shift which resulted in 0.45, -0.29, and 0.56 mean Pearson correlation, respectively. FDD could be a valuable tool for care providers and vendors who need to verify if a MIL system is likely to perform reliably when implemented at a new site, without requiring any additional annotations from pathologists.

Previous unsupervised anomaly detection (UAD) methods often struggle with significant intra-class diversity; i.e., a class in a dataset contains multiple subclasses, which we categorize as Feature-Rich Anomaly Detection Datasets (FRADs). This is evident in applications such as unified setting and unmanned supermarket scenarios. To address this challenge, we developed MiniMaxAD: a lightweight autoencoder designed to efficiently compress and memorize extensive information from normal images. Our model utilizes a large kernel convolutional network equipped with a Global Response Normalization (GRN) unit and employs a multi-scale feature reconstruction strategy. The GRN unit significantly increases the upper limit of the network's capacity, while the large kernel convolution facilitates the extraction of highly abstract patterns, leading to compact normal feature modeling. Additionally, we introduce an Adaptive Contraction Loss (ADCLoss), tailored to FRADs to overcome the limitations of global cosine distance loss. MiniMaxAD was comprehensively tested across six challenging UAD benchmarks, achieving state-of-the-art results in four and highly competitive outcomes in the remaining two. Notably, our model achieved a detection AUROC of up to 97.0\% in ViSA under the unified setting. Moreover, it not only achieved state-of-the-art performance in unmanned supermarket tasks but also exhibited an inference speed 37 times faster than the previous best method, demonstrating its effectiveness in complex UAD tasks.

Most existing attention prediction research focuses on salient instances like humans and objects. However, the more complex interaction-oriented attention, arising from the comprehension of interactions between instances by human observers, remains largely unexplored. This is equally crucial for advancing human-machine interaction and human-centered artificial intelligence. To bridge this gap, we first collect a novel gaze fixation dataset named IG, comprising 530,000 fixation points across 740 diverse interaction categories, capturing visual attention during human observers cognitive processes of interactions. Subsequently, we introduce the zero-shot interaction-oriented attention prediction task ZeroIA, which challenges models to predict visual cues for interactions not encountered during training. Thirdly, we present the Interactive Attention model IA, designed to emulate human observers cognitive processes to tackle the ZeroIA problem. Extensive experiments demonstrate that the proposed IA outperforms other state-of-the-art approaches in both ZeroIA and fully supervised settings. Lastly, we endeavor to apply interaction-oriented attention to the interaction recognition task itself. Further experimental results demonstrate the promising potential to enhance the performance and interpretability of existing state-of-the-art HOI models by incorporating real human attention data from IG and attention labels generated by IA.

Infrared physical adversarial examples are of great significance for studying the security of infrared AI systems that are widely used in our lives such as autonomous driving. Previous infrared physical attacks mainly focused on 2D infrared pedestrian detection which may not fully manifest its destructiveness to AI systems. In this work, we propose a physical attack method against infrared detectors based on 3D modeling, which is applied to a real car. The goal is to design a set of infrared adversarial stickers to make cars invisible to infrared detectors at various viewing angles, distances, and scenes. We build a 3D infrared car model with real infrared characteristics and propose an infrared adversarial pattern generation method based on 3D mesh shadow. We propose a 3D control points-based mesh smoothing algorithm and use a set of smoothness loss functions to enhance the smoothness of adversarial meshes and facilitate the sticker implementation. Besides, We designed the aluminum stickers and conducted physical experiments on two real Mercedes-Benz A200L cars. Our adversarial stickers hid the cars from Faster RCNN, an object detector, at various viewing angles, distances, and scenes. The attack success rate (ASR) was 91.49% for real cars. In comparison, the ASRs of random stickers and no sticker were only 6.21% and 0.66%, respectively. In addition, the ASRs of the designed stickers against six unseen object detectors such as YOLOv3 and Deformable DETR were between 73.35%-95.80%, showing good transferability of the attack performance across detectors.

In this paper, we review the NTIRE 2024 challenge on Restore Any Image Model (RAIM) in the Wild. The RAIM challenge constructed a benchmark for image restoration in the wild, including real-world images with/without reference ground truth in various scenarios from real applications. The participants were required to restore the real-captured images from complex and unknown degradation, where generative perceptual quality and fidelity are desired in the restoration result. The challenge consisted of two tasks. Task one employed real referenced data pairs, where quantitative evaluation is available. Task two used unpaired images, and a comprehensive user study was conducted. The challenge attracted more than 200 registrations, where 39 of them submitted results with more than 400 submissions. Top-ranked methods improved the state-of-the-art restoration performance and obtained unanimous recognition from all 18 judges. The proposed datasets are available at https://drive.google.com/file/d/1DqbxUoiUqkAIkExu3jZAqoElr_nu1IXb/view?usp=sharing and the homepage of this challenge is at https://codalab.lisn.upsaclay.fr/competitions/17632.

Large-scale "foundation models" have gained traction as a way to leverage the vast amounts of unlabeled remote sensing data collected every day. However, due to the multiplicity of Earth Observation satellites, these models should learn "sensor agnostic" representations, that generalize across sensor characteristics with minimal fine-tuning. This is complicated by data availability, as low-resolution imagery, such as Sentinel-2 and Landsat-8 data, are available in large amounts, while very high-resolution aerial or satellite data is less common. To tackle these challenges, we introduce cross-sensor self-supervised training and alignment for remote sensing (X-STARS). We design a self-supervised training loss, the Multi-Sensor Alignment Dense loss (MSAD), to align representations across sensors, even with vastly different resolutions. Our X-STARS can be applied to train models from scratch, or to adapt large models pretrained on e.g low-resolution EO data to new high-resolution sensors, in a continual pretraining framework. We collect and release MSC-France, a new multi-sensor dataset, on which we train our X-STARS models, then evaluated on seven downstream classification and segmentation tasks. We demonstrate that X-STARS outperforms the state-of-the-art by a significant margin with less data across various conditions of data availability and resolutions.

Encryption on the internet with the shift to HTTPS has been an important step to improve the privacy of internet users. However, there is an increasing body of work about extracting information from encrypted internet traffic without having to decrypt it. Such attacks bypass security guarantees assumed to be given by HTTPS and thus need to be understood. Prior works showed that the variable bitrates of video streams are sufficient to identify which video someone is watching. These works generally have to make trade-offs in aspects such as accuracy, scalability, robustness, etc. These trade-offs complicate the practical use of these attacks. To that end, we propose a deep metric learning framework based on the triplet loss method. Through this framework, we achieve robust, generalisable, scalable and transferable encrypted video stream detection. First, the triplet loss is better able to deal with video streams not seen during training. Second, our approach can accurately classify videos not seen during training. Third, we show that our method scales well to a dataset of over 1000 videos. Finally, we show that a model trained on video streams over Chrome can also classify streams over Firefox. Our results suggest that this side-channel attack is more broadly applicable than originally thought. We provide our code alongside a diverse and up-to-date dataset for future research.

We introduce RoScenes, the largest multi-view roadside perception dataset, which aims to shed light on the development of vision-centric Bird's Eye View (BEV) approaches for more challenging traffic scenes. The highlights of RoScenes include significantly large perception area, full scene coverage and crowded traffic. More specifically, our dataset achieves surprising 21.13M 3D annotations within 64,000 $m^2$. To relieve the expensive costs of roadside 3D labeling, we present a novel BEV-to-3D joint annotation pipeline to efficiently collect such a large volume of data. After that, we organize a comprehensive study for current BEV methods on RoScenes in terms of effectiveness and efficiency. Tested methods suffer from the vast perception area and variation of sensor layout across scenes, resulting in performance levels falling below expectations. To this end, we propose RoBEV that incorporates feature-guided position embedding for effective 2D-3D feature assignment. With its help, our method outperforms state-of-the-art by a large margin without extra computational overhead on validation set. Our dataset and devkit will be made available at \url{https://github.com/xiaosu-zhu/RoScenes}.

With the rapid development of face recognition (FR) systems, the privacy of face images on social media is facing severe challenges due to the abuse of unauthorized FR systems. Some studies utilize adversarial attack techniques to defend against malicious FR systems by generating adversarial examples. However, the generated adversarial examples, i.e., the protected face images, tend to suffer from subpar visual quality and low transferability. In this paper, we propose a novel face protection approach, dubbed DiffAM, which leverages the powerful generative ability of diffusion models to generate high-quality protected face images with adversarial makeup transferred from reference images. To be specific, we first introduce a makeup removal module to generate non-makeup images utilizing a fine-tuned diffusion model with guidance of textual prompts in CLIP space. As the inverse process of makeup transfer, makeup removal can make it easier to establish the deterministic relationship between makeup domain and non-makeup domain regardless of elaborate text prompts. Then, with this relationship, a CLIP-based makeup loss along with an ensemble attack strategy is introduced to jointly guide the direction of adversarial makeup domain, achieving the generation of protected face images with natural-looking makeup and high black-box transferability. Extensive experiments demonstrate that DiffAM achieves higher visual quality and attack success rates with a gain of 12.98% under black-box setting compared with the state of the arts. The code will be available at https://github.com/HansSunY/DiffAM.

3D face registration is an important process in which a 3D face model is aligned and mapped to a template face. However, the task of 3D face registration becomes particularly challenging when dealing with partial face data, where only limited facial information is available. To address this challenge, this paper presents a novel deep learning-based approach that combines quasi-conformal geometry with deep neural networks for partial face registration. The proposed framework begins with a Landmark Detection Network that utilizes curvature information to detect the presence of facial features and estimate their corresponding coordinates. These facial landmark features serve as essential guidance for the registration process. To establish a dense correspondence between the partial face and the template surface, a registration network based on quasiconformal theories is employed. The registration network establishes a bijective quasiconformal surface mapping aligning corresponding partial faces based on detected landmarks and curvature values. It consists of the Coefficients Prediction Network, which outputs the optimal Beltrami coefficient representing the surface mapping. The Beltrami coefficient quantifies the local geometric distortion of the mapping. By controlling the magnitude of the Beltrami coefficient through a suitable activation function, the bijectivity and geometric distortion of the mapping can be controlled. The Beltrami coefficient is then fed into the Beltrami solver network to reconstruct the corresponding mapping. The surface registration enables the acquisition of corresponding regions and the establishment of point-wise correspondence between different partial faces, facilitating precise shape comparison through the evaluation of point-wise geometric differences at these corresponding regions. Experimental results demonstrate the effectiveness of the proposed method.

Recent advances in generative models trained on large-scale datasets have made it possible to synthesize high-quality samples across various domains. Moreover, the emergence of strong inversion networks enables not only a reconstruction of real-world images but also the modification of attributes through various editing methods. However, in certain domains related to privacy issues, e.g., human faces, advanced generative models along with strong inversion methods can lead to potential misuses. In this paper, we propose an essential yet under-explored task called generative identity unlearning, which steers the model not to generate an image of a specific identity. In the generative identity unlearning, we target the following objectives: (i) preventing the generation of images with a certain identity, and (ii) preserving the overall quality of the generative model. To satisfy these goals, we propose a novel framework, Generative Unlearning for Any Identity (GUIDE), which prevents the reconstruction of a specific identity by unlearning the generator with only a single image. GUIDE consists of two parts: (i) finding a target point for optimization that un-identifies the source latent code and (ii) novel loss functions that facilitate the unlearning procedure while less affecting the learned distribution. Our extensive experiments demonstrate that our proposed method achieves state-of-the-art performance in the generative machine unlearning task. The code is available at https://github.com/KHU-AGI/GUIDE.

We present Dual3D, a novel text-to-3D generation framework that generates high-quality 3D assets from texts in only $1$ minute.The key component is a dual-mode multi-view latent diffusion model. Given the noisy multi-view latents, the 2D mode can efficiently denoise them with a single latent denoising network, while the 3D mode can generate a tri-plane neural surface for consistent rendering-based denoising. Most modules for both modes are tuned from a pre-trained text-to-image latent diffusion model to circumvent the expensive cost of training from scratch. To overcome the high rendering cost during inference, we propose the dual-mode toggling inference strategy to use only $1/10$ denoising steps with 3D mode, successfully generating a 3D asset in just $10$ seconds without sacrificing quality. The texture of the 3D asset can be further enhanced by our efficient texture refinement process in a short time. Extensive experiments demonstrate that our method delivers state-of-the-art performance while significantly reducing generation time. Our project page is available at https://dual3d.github.io

Infrared (IR) image super-resolution faces challenges from homogeneous background pixel distributions and sparse target regions, requiring models that effectively handle long-range dependencies and capture detailed local-global information. Recent advancements in Mamba-based (Selective Structured State Space Model) models, employing state space models, have shown significant potential in visual tasks, suggesting their applicability for IR enhancement. In this work, we introduce IRSRMamba: Infrared Image Super-Resolution via Mamba-based Wavelet Transform Feature Modulation Model, a novel Mamba-based model designed specifically for IR image super-resolution. This model enhances the restoration of context-sparse target details through its advanced dependency modeling capabilities. Additionally, a new wavelet transform feature modulation block improves multi-scale receptive field representation, capturing both global and local information efficiently. Comprehensive evaluations confirm that IRSRMamba outperforms existing models on multiple benchmarks. This research advances IR super-resolution and demonstrates the potential of Mamba-based models in IR image processing. Code are available at \url{https://github.com/yongsongH/IRSRMamba}.

The study of astronomical phenomena through ground-based observations is always challenged by the distorting effects of Earth's atmosphere. Traditional methods of post-facto image correction, essential for correcting these distortions, often rely on simplifying assumptions that limit their effectiveness, particularly in the presence of spatially variant atmospheric turbulence. Such cases are often solved by partitioning the field-of-view into small patches, deconvolving each patch independently, and merging all patches together. This approach is often inefficient and can produce artifacts. Recent advancements in computational techniques and the advent of deep learning offer new pathways to address these limitations. This paper introduces a novel framework leveraging a deep neural network to emulate spatially variant convolutions, offering a breakthrough in the efficiency and accuracy of astronomical image deconvolution. By training on a dataset of images convolved with spatially invariant point spread functions and validating its generalizability to spatially variant conditions, this approach presents a significant advancement over traditional methods. The convolution emulator is used as a forward model in a multi-object multi-frame blind deconvolution algorithm for solar images. The emulator enables the deconvolution of solar observations across large fields of view without resorting to patch-wise mosaicking, thus avoiding artifacts associated with such techniques. This method represents a significant computational advantage, reducing processing times by orders of magnitude.

Box-free model watermarking is an emerging technique to safeguard the intellectual property of deep learning models, particularly those for low-level image processing tasks. Existing works have verified and improved its effectiveness in several aspects. However, in this paper, we reveal that box-free model watermarking is prone to removal attacks, even under the real-world threat model such that the protected model and the watermark extractor are in black boxes. Under this setting, we carry out three studies. 1) We develop an extractor-gradient-guided (EGG) remover and show its effectiveness when the extractor uses ReLU activation only. 2) More generally, for an unknown extractor, we leverage adversarial attacks and design the EGG remover based on the estimated gradients. 3) Under the most stringent condition that the extractor is inaccessible, we design a transferable remover based on a set of private proxy models. In all cases, the proposed removers can successfully remove embedded watermarks while preserving the quality of the processed images, and we also demonstrate that the EGG remover can even replace the watermarks. Extensive experimental results verify the effectiveness and generalizability of the proposed attacks, revealing the vulnerabilities of the existing box-free methods and calling for further research.

This paper addresses the unrealistic aspect of the commonly adopted Continuous Incremental Semantic Segmentation (CISS) scenario, termed overlapped. We point out that overlapped allows the same image to reappear in future tasks with different pixel labels, which is far from practical incremental learning scenarios. Moreover, we identified that this flawed scenario may lead to biased results for two commonly used techniques in CISS, pseudo-labeling and exemplar memory, resulting in unintended advantages or disadvantages for certain techniques. To mitigate this, a practical scenario called partitioned is proposed, in which the dataset is first divided into distinct subsets representing each class, and then the subsets are assigned to each corresponding task. This efficiently addresses the issue above while meeting the requirement of CISS scenario, such as capturing the background shifts. Furthermore, we identify and address the code implementation issues related to retrieving data from the exemplar memory, which was ignored in previous works. Lastly, we introduce a simple yet competitive memory-based baseline, MiB-AugM, that handles background shifts of current tasks in the exemplar memory. This baseline achieves state-of-the-art results across multiple tasks involving learning numerous new classes.

Automated region of interest detection in histopathological image analysis is a challenging and important topic with tremendous potential impact on clinical practice. The deep-learning methods used in computational pathology may help us to reduce costs and increase the speed and accuracy of cancer diagnosis. We started with the UNC Melanocytic Tumor Dataset cohort that contains 160 hematoxylin and eosin whole-slide images of primary melanomas (86) and nevi (74). We randomly assigned 80% (134) as a training set and built an in-house deep-learning method to allow for classification, at the slide level, of nevi and melanomas. The proposed method performed well on the other 20% (26) test dataset; the accuracy of the slide classification task was 92.3% and our model also performed well in terms of predicting the region of interest annotated by the pathologists, showing excellent performance of our model on melanocytic skin tumors. Even though we tested the experiments on the skin tumor dataset, our work could also be extended to other medical image detection problems to benefit the clinical evaluation and diagnosis of different tumors.

Multi-line LiDAR is widely used in autonomous vehicles, so point cloud-based 3D detectors are essential for autonomous driving. Extracting rich multi-scale features is crucial for point cloud-based 3D detectors in autonomous driving due to significant differences in the size of different types of objects. However, due to the real-time requirements, large-size convolution kernels are rarely used to extract large-scale features in the backbone. Current 3D detectors commonly use feature pyramid networks to obtain large-scale features; however, some objects containing fewer point clouds are further lost during downsampling, resulting in degraded performance. Since pillar-based schemes require much less computation than voxel-based schemes, they are more suitable for constructing real-time 3D detectors. Hence, we propose PillarNeXt, a pillar-based scheme. We redesigned the feature encoding, the backbone, and the neck of the 3D detector. We propose Voxel2Pillar feature encoding, which uses a sparse convolution constructor to construct pillars with richer point cloud features, especially height features. Moreover, additional learnable parameters are added, which enables the initial pillar to achieve higher performance capabilities. We extract multi-scale and large-scale features in the proposed fully sparse backbone, which does not utilize large-size convolutional kernels; the backbone consists of the proposed multi-scale feature extraction module. The neck consists of the proposed sparse ConvNeXt, whose simple structure significantly improves the performance. The effectiveness of the proposed PillarNeXt is validated on the Waymo Open Dataset, and object detection accuracy for vehicles, pedestrians, and cyclists is improved; we also verify the effectiveness of each proposed module in detail.

High-level visual brain regions contain subareas in which neurons appear to respond more strongly to examples of a particular semantic category, like faces or bodies, rather than objects. However, recent work has shown that while this finding holds on average, some out-of-category stimuli also activate neurons in these regions. This may be due to visual features common among the preferred class also being present in other images. Here, we propose a deep-learning-based approach for visualizing these features. For each neuron, we identify relevant visual features driving its selectivity by modelling responses to images based on latent activations of a deep neural network. Given an out-of-category image which strongly activates the neuron, our method first identifies a reference image from the preferred category yielding a similar feature activation pattern. We then backpropagate latent activations of both images to the pixel level, while enhancing the identified shared dimensions and attenuating non-shared features. The procedure highlights image regions containing shared features driving responses of the model neuron. We apply the algorithm to novel recordings from body-selective regions in macaque IT cortex in order to understand why some images of objects excite these neurons. Visualizations reveal object parts which resemble parts of a macaque body, shedding light on neural preference of these objects.

Continual learning, involving sequential training on diverse tasks, often faces catastrophic forgetting. While knowledge distillation-based approaches exhibit notable success in preventing forgetting, we pinpoint a limitation in their ability to distill the cumulative knowledge of all the previous tasks. To remedy this, we propose Dense Knowledge Distillation (DKD). DKD uses a task pool to track the model's capabilities. It partitions the output logits of the model into dense groups, each corresponding to a task in the task pool. It then distills all tasks' knowledge using all groups. However, using all the groups can be computationally expensive, we also suggest random group selection in each optimization step. Moreover, we propose an adaptive weighting scheme, which balances the learning of new classes and the retention of old classes, based on the count and similarity of the classes. Our DKD outperforms recent state-of-the-art baselines across diverse benchmarks and scenarios. Empirical analysis underscores DKD's ability to enhance model stability, promote flatter minima for improved generalization, and remains robust across various memory budgets and task orders. Moreover, it seamlessly integrates with other CL methods to boost performance and proves versatile in offline scenarios like model compression.

In visual tasks, large teacher models capture essential features and deep information, enhancing performance. However, distilling this information into smaller student models often leads to performance loss due to structural differences and capacity limitations. To tackle this, we propose a distillation framework based on graph knowledge, including a multi-level feature alignment strategy and an attention-guided mechanism to provide a targeted learning trajectory for the student model. We emphasize spectral embedding (SE) as a key technique in our distillation process, which merges the student's feature space with the relational knowledge and structural complexities similar to the teacher network. This method captures the teacher's understanding in a graph-based representation, enabling the student model to more accurately mimic the complex structural dependencies present in the teacher model. Compared to methods that focus only on specific distillation areas, our strategy not only considers key features within the teacher model but also endeavors to capture the relationships and interactions among feature sets, encoding these complex pieces of information into a graph structure to understand and utilize the dynamic relationships among these pieces of information from a global perspective. Experiments show that our method outperforms previous feature distillation methods on the CIFAR-100, MS-COCO, and Pascal VOC datasets, proving its efficiency and applicability.

In this work, we present Semantic Gesticulator, a novel framework designed to synthesize realistic gestures accompanying speech with strong semantic correspondence. Semantically meaningful gestures are crucial for effective non-verbal communication, but such gestures often fall within the long tail of the distribution of natural human motion. The sparsity of these movements makes it challenging for deep learning-based systems, trained on moderately sized datasets, to capture the relationship between the movements and the corresponding speech semantics. To address this challenge, we develop a generative retrieval framework based on a large language model. This framework efficiently retrieves suitable semantic gesture candidates from a motion library in response to the input speech. To construct this motion library, we summarize a comprehensive list of commonly used semantic gestures based on findings in linguistics, and we collect a high-quality motion dataset encompassing both body and hand movements. We also design a novel GPT-based model with strong generalization capabilities to audio, capable of generating high-quality gestures that match the rhythm of speech. Furthermore, we propose a semantic alignment mechanism to efficiently align the retrieved semantic gestures with the GPT's output, ensuring the naturalness of the final animation. Our system demonstrates robustness in generating gestures that are rhythmically coherent and semantically explicit, as evidenced by a comprehensive collection of examples. User studies confirm the quality and human-likeness of our results, and show that our system outperforms state-of-the-art systems in terms of semantic appropriateness by a clear margin.

Diffusion models have recently gained significant traction due to their ability to generate high-fidelity and diverse images and videos conditioned on text prompts. In medicine, this application promises to address the critical challenge of data scarcity, a consequence of barriers in data sharing, stringent patient privacy regulations, and disparities in patient population and demographics. By generating realistic and varying medical 2D and 3D images, these models offer a rich, privacy-respecting resource for algorithmic training and research. To this end, we introduce MediSyn, a pair of instruction-tuned text-guided latent diffusion models with the ability to generate high-fidelity and diverse medical 2D and 3D images across specialties and modalities. Through established metrics, we show significant improvement in broad medical image and video synthesis guided by text prompts.

Large language models are well-known to be effective at few-shot in-context learning (ICL). Recent advancements in multimodal foundation models have enabled unprecedentedly long context windows, presenting an opportunity to explore their capability to perform ICL with many more demonstrating examples. In this work, we evaluate the performance of multimodal foundation models scaling from few-shot to many-shot ICL. We benchmark GPT-4o and Gemini 1.5 Pro across 10 datasets spanning multiple domains (natural imagery, medical imagery, remote sensing, and molecular imagery) and tasks (multi-class, multi-label, and fine-grained classification). We observe that many-shot ICL, including up to almost 2,000 multimodal demonstrating examples, leads to substantial improvements compared to few-shot (<100 examples) ICL across all of the datasets. Further, Gemini 1.5 Pro performance continues to improve log-linearly up to the maximum number of tested examples on many datasets. Given the high inference costs associated with the long prompts required for many-shot ICL, we also explore the impact of batching multiple queries in a single API call. We show that batching up to 50 queries can lead to performance improvements under zero-shot and many-shot ICL, with substantial gains in the zero-shot setting on multiple datasets, while drastically reducing per-query cost and latency. Finally, we measure ICL data efficiency of the models, or the rate at which the models learn from more demonstrating examples. We find that while GPT-4o and Gemini 1.5 Pro achieve similar zero-shot performance across the datasets, Gemini 1.5 Pro exhibits higher ICL data efficiency than GPT-4o on most datasets. Our results suggest that many-shot ICL could enable users to efficiently adapt multimodal foundation models to new applications and domains. Our codebase is publicly available at https://github.com/stanfordmlgroup/ManyICL .

Due to spatial redundancy in remote sensing images, sparse tokens containing rich information are usually involved in self-attention (SA) to reduce the overall token numbers within the calculation, avoiding the high computational cost issue in Vision Transformers. However, such methods usually obtain sparse tokens by hand-crafted or parallel-unfriendly designs, posing a challenge to reach a better balance between efficiency and performance. Different from them, this paper proposes to use learnable meta tokens to formulate sparse tokens, which effectively learn key information meanwhile improving the inference speed. Technically, the meta tokens are first initialized from image tokens via cross-attention. Then, we propose Dual Cross-Attention (DCA) to promote information exchange between image tokens and meta tokens, where they serve as query and key (value) tokens alternatively in a dual-branch structure, significantly reducing the computational complexity compared to self-attention. By employing DCA in the early stages with dense visual tokens, we obtain the hierarchical architecture LeMeViT with various sizes. Experimental results in classification and dense prediction tasks show that LeMeViT has a significant $1.7 \times$ speedup, fewer parameters, and competitive performance compared to the baseline models, and achieves a better trade-off between efficiency and performance.

We describe the design and results from the BraTS 2023 Intracranial Meningioma Segmentation Challenge. The BraTS Meningioma Challenge differed from prior BraTS Glioma challenges in that it focused on meningiomas, which are typically benign extra-axial tumors with diverse radiologic and anatomical presentation and a propensity for multiplicity. Nine participating teams each developed deep-learning automated segmentation models using image data from the largest multi-institutional systematically expert annotated multilabel multi-sequence meningioma MRI dataset to date, which included 1000 training set cases, 141 validation set cases, and 283 hidden test set cases. Each case included T2, T2/FLAIR, T1, and T1Gd brain MRI sequences with associated tumor compartment labels delineating enhancing tumor, non-enhancing tumor, and surrounding non-enhancing T2/FLAIR hyperintensity. Participant automated segmentation models were evaluated and ranked based on a scoring system evaluating lesion-wise metrics including dice similarity coefficient (DSC) and 95% Hausdorff Distance. The top ranked team had a lesion-wise median dice similarity coefficient (DSC) of 0.976, 0.976, and 0.964 for enhancing tumor, tumor core, and whole tumor, respectively and a corresponding average DSC of 0.899, 0.904, and 0.871, respectively. These results serve as state-of-the-art benchmarks for future pre-operative meningioma automated segmentation algorithms. Additionally, we found that 1286 of 1424 cases (90.3%) had at least 1 compartment voxel abutting the edge of the skull-stripped image edge, which requires further investigation into optimal pre-processing face anonymization steps.

This paper explores the size-invariance of evaluation metrics in Salient Object Detection (SOD), especially when multiple targets of diverse sizes co-exist in the same image. We observe that current metrics are size-sensitive, where larger objects are focused, and smaller ones tend to be ignored. We argue that the evaluation should be size-invariant because bias based on size is unjustified without additional semantic information. In pursuit of this, we propose a generic approach that evaluates each salient object separately and then combines the results, effectively alleviating the imbalance. We further develop an optimization framework tailored to this goal, achieving considerable improvements in detecting objects of different sizes. Theoretically, we provide evidence supporting the validity of our new metrics and present the generalization analysis of SOD. Extensive experiments demonstrate the effectiveness of our method. The code is available at https://github.com/Ferry-Li/SI-SOD.

This paper investigates an extremely challenging problem, barely-supervised medical image segmentation (BSS), where the training dataset comprises limited labeled data with only single-slice annotations and numerous unlabeled images. Currently, state-of-the-art (SOTA) BSS methods utilize a registration-based paradigm, depending on image registration to propagate single-slice annotations into volumetric pseudo labels for constructing a complete labeled set. However, this paradigm has a critical limitation: the pseudo labels generated by image registration are unreliable and noisy. Motivated by this, we propose a new perspective: training a model using only single-annotated slices as the labeled set without relying on image registration. To this end, we formulate BSS as an unsupervised domain adaptation (UDA) problem. Specifically, we first design a novel noise-free labeled data construction algorithm (NFC) for slice-to-volume labeled data synthesis, which may result in a side effect: domain shifts between the synthesized images and the original images. Then, a frequency and spatial mix-up strategy (FSX) is further introduced to mitigate the domain shifts for UDA. Extensive experiments demonstrate that our method provides a promising alternative for BSS. Remarkably, the proposed method with only one labeled slice achieves an 80.77% dice score on left atrial segmentation, outperforming the SOTA by 61.28%. The code will be released upon the publication of this paper.

3D cameras have emerged as a critical source of information for applications in robotics and autonomous driving. These cameras provide robots with the ability to capture and utilize point clouds, enabling them to navigate their surroundings and avoid collisions with other objects. However, current standard camera evaluation metrics often fail to consider the specific application context. These metrics typically focus on measures like Chamfer distance (CD) or Earth Mover's Distance (EMD), which may not directly translate to performance in real-world scenarios. To address this limitation, we propose a novel metric for point cloud evaluation, specifically designed to assess the suitability of 3D cameras for the critical task of collision avoidance. This metric incorporates application-specific considerations and provides a more accurate measure of a camera's effectiveness in ensuring safe robot navigation.

For robotics applications where there is a limited number of (typically ego-centric) views, parametric representations such as neural radiance fields (NeRFs) generalize better than non-parametric ones such as Gaussian splatting (GS) to views that are very different from those in the training data; GS however can render much faster than NeRFs. We develop a procedure to convert back and forth between the two. Our approach achieves the best of both NeRFs (superior PSNR, SSIM, and LPIPS on dissimilar views, and a compact representation) and GS (real-time rendering and ability for easily modifying the representation); the computational cost of these conversions is minor compared to training the two from scratch.

The advances in deep generative models have greatly accelerate the process of video procession such as video enhancement and synthesis. Learning spatio-temporal video models requires to capture the temporal dynamics of a scene, in addition to the visual appearance of individual frames. Illumination consistency, which reflects the variations of illumination in the dynamic video sequences, play a vital role in video processing. Unfortunately, to date, no well-accepted quantitative metric has been proposed for video illumination consistency evaluation. In this paper, we propose a illumination histogram consistency (IHC) metric to quantitatively and automatically evaluate the illumination consistency of the video sequences. IHC measures the illumination variation of any video sequence based on the illumination histogram discrepancies across all the frames in the video sequence. Specifically, given a video sequence, we first estimate the illumination map of each individual frame using the Retinex model; Then, using the illumination maps, the mean illumination histogram of the video sequence is computed by the mean operation across all the frames; Next, we compute the illumination histogram discrepancy between each individual frame and the mean illumination histogram and sum up all the illumination histogram discrepancies to represent the illumination variations of the video sequence. Finally, we obtain the IHC score from the illumination histogram discrepancies via normalization and subtraction operations. Experiments are conducted to illustrate the performance of the proposed IHC metric and its capability to measure the illumination variations in video sequences. The source code is available on \url{https://github.com/LongChenCV/IHC-Metric}.

Learning commonsense reasoning from visual contexts and scenes in real-world is a crucial step toward advanced artificial intelligence. However, existing video reasoning benchmarks are still inadequate since they were mainly designed for factual or situated reasoning and rarely involve broader knowledge in the real world. Our work aims to delve deeper into reasoning evaluations, specifically within dynamic, open-world, and structured context knowledge. We propose a new benchmark (SOK-Bench), consisting of 44K questions and 10K situations with instance-level annotations depicted in the videos. The reasoning process is required to understand and apply situated knowledge and general knowledge for problem-solving. To create such a dataset, we propose an automatic and scalable generation method to generate question-answer pairs, knowledge graphs, and rationales by instructing the combinations of LLMs and MLLMs. Concretely, we first extract observable situated entities, relations, and processes from videos for situated knowledge and then extend to open-world knowledge beyond the visible content. The task generation is facilitated through multiple dialogues as iterations and subsequently corrected and refined by our designed self-promptings and demonstrations. With a corpus of both explicit situated facts and implicit commonsense, we generate associated question-answer pairs and reasoning processes, finally followed by manual reviews for quality assurance. We evaluated recent mainstream large vision-language models on the benchmark and found several insightful conclusions. For more information, please refer to our benchmark at www.bobbywu.com/SOKBench.

Reasoning in the real world is not divorced from situations. How to capture the present knowledge from surrounding situations and perform reasoning accordingly is crucial and challenging for machine intelligence. This paper introduces a new benchmark that evaluates the situated reasoning ability via situation abstraction and logic-grounded question answering for real-world videos, called Situated Reasoning in Real-World Videos (STAR Benchmark). This benchmark is built upon the real-world videos associated with human actions or interactions, which are naturally dynamic, compositional, and logical. The dataset includes four types of questions, including interaction, sequence, prediction, and feasibility. We represent the situations in real-world videos by hyper-graphs connecting extracted atomic entities and relations (e.g., actions, persons, objects, and relationships). Besides visual perception, situated reasoning also requires structured situation comprehension and logical reasoning. Questions and answers are procedurally generated. The answering logic of each question is represented by a functional program based on a situation hyper-graph. We compare various existing video reasoning models and find that they all struggle on this challenging situated reasoning task. We further propose a diagnostic neuro-symbolic model that can disentangle visual perception, situation abstraction, language understanding, and functional reasoning to understand the challenges of this benchmark.

Correspondence-based statistical shape modeling (SSM) stands as a powerful technology for morphometric analysis in clinical research. SSM facilitates population-level characterization and quantification of anatomical shapes such as bones and organs, aiding in pathology and disease diagnostics and treatment planning. Despite its potential, SSM remains under-utilized in medical research due to the significant overhead associated with automatic construction methods, which demand complete, aligned shape surface representations. Additionally, optimization-based techniques rely on bias-inducing assumptions or templates and have prolonged inference times as the entire cohort is simultaneously optimized. To overcome these challenges, we introduce Point2SSM++, a principled, self-supervised deep learning approach that directly learns correspondence points from point cloud representations of anatomical shapes. Point2SSM++ is robust to misaligned and inconsistent input, providing SSM that accurately samples individual shape surfaces while effectively capturing population-level statistics. Additionally, we present principled extensions of Point2SSM++ to adapt it for dynamic spatiotemporal and multi-anatomy use cases, demonstrating the broad versatility of the Point2SSM++ framework. Furthermore, we present extensions of Point2SSM++ tailored for dynamic spatiotemporal and multi-anatomy scenarios, showcasing the broad versatility of the framework. Through extensive validation across diverse anatomies, evaluation metrics, and clinically relevant downstream tasks, we demonstrate Point2SSM++'s superiority over existing state-of-the-art deep learning models and traditional approaches. Point2SSM++ substantially enhances the feasibility of SSM generation and significantly broadens its array of potential clinical applications.

Anatomical shape analysis plays a pivotal role in clinical research and hypothesis testing, where the relationship between form and function is paramount. Correspondence-based statistical shape modeling (SSM) facilitates population-level morphometrics but requires a cumbersome, potentially bias-inducing construction pipeline. Recent advancements in deep learning have streamlined this process in inference by providing SSM prediction directly from unsegmented medical images. However, the proposed approaches are fully supervised and require utilizing a traditional SSM construction pipeline to create training data, thus inheriting the associated burdens and limitations. To address these challenges, we introduce a weakly supervised deep learning approach to predict SSM from images using point cloud supervision. Specifically, we propose reducing the supervision associated with the state-of-the-art fully Bayesian variational information bottleneck DeepSSM (BVIB-DeepSSM) model. BVIB-DeepSSM is an effective, principled framework for predicting probabilistic anatomical shapes from images with quantification of both aleatoric and epistemic uncertainties. Whereas the original BVIB-DeepSSM method requires strong supervision in the form of ground truth correspondence points, the proposed approach utilizes weak supervision via point cloud surface representations, which are more readily obtainable. Furthermore, the proposed approach learns correspondence in a completely data-driven manner without prior assumptions about the expected variability in shape cohort. Our experiments demonstrate that this approach yields similar accuracy and uncertainty estimation to the fully supervised scenario while substantially enhancing the feasibility of model training for SSM construction.

This study examines the role of visual highlights in guiding user attention in drone monitoring tasks, employing a simulated interface for observation. The experiment results show that such highlights can significantly expedite the visual attention on the corresponding area. Based on this observation, we leverage both the temporal and spatial information in the highlight to develop a new saliency model: the highlight-informed saliency model (HISM), to infer the visual attention change in the highlight condition. Our findings show the effectiveness of visual highlights in enhancing user attention and demonstrate the potential of incorporating these cues into saliency prediction models.

Unsupervised Domain Adaptation (UDA) aims to transfer knowledge learned from a labeled source domain to an unlabeled target domain. While UDA methods for synthetic to real-world domains (synth-to-real) show remarkable performance in tasks such as semantic segmentation and object detection, very few were proposed for the instance segmentation task. In this paper, we introduce UDA4Inst, a model of synth-to-real UDA for instance segmentation in autonomous driving. We propose a novel cross-domain bidirectional data mixing method at the instance level to fully leverage the data from both source and target domains. Rare-class balancing and category module training are also employed to further improve the performance. It is worth noting that we are the first to demonstrate results on two new synth-to-real instance segmentation benchmarks, with 39.0 mAP on UrbanSyn->Cityscapes and 35.7 mAP on Synscapes->Cityscapes. UDA4Inst also achieves the state-of-the-art result on SYNTHIA->Cityscapes with 31.3 mAP, +15.6 higher than the latest approach. Our code will be released.

The systematic evaluation and understanding of computer vision models under varying conditions require large amounts of data with comprehensive and customized labels, which real-world vision datasets rarely satisfy. While current synthetic data generators offer a promising alternative, particularly for embodied AI tasks, they often fall short for computer vision tasks due to low asset and rendering quality, limited diversity, and unrealistic physical properties. We introduce the BEHAVIOR Vision Suite (BVS), a set of tools and assets to generate fully customized synthetic data for systematic evaluation of computer vision models, based on the newly developed embodied AI benchmark, BEHAVIOR-1K. BVS supports a large number of adjustable parameters at the scene level (e.g., lighting, object placement), the object level (e.g., joint configuration, attributes such as "filled" and "folded"), and the camera level (e.g., field of view, focal length). Researchers can arbitrarily vary these parameters during data generation to perform controlled experiments. We showcase three example application scenarios: systematically evaluating the robustness of models across different continuous axes of domain shift, evaluating scene understanding models on the same set of images, and training and evaluating simulation-to-real transfer for a novel vision task: unary and binary state prediction. Project website: https://behavior-vision-suite.github.io/

Aerial imagery is increasingly used in Earth science and natural resource management as a complement to labor-intensive ground-based surveys. Aerial systems can collect overlapping images that provide multiple views of each location from different perspectives. However, most prediction approaches (e.g. for tree species classification) use a single, synthesized top-down "orthomosaic" image as input that contains little to no information about the vertical aspects of objects and may include processing artifacts. We propose an alternate approach that generates predictions directly on the raw images and accurately maps these predictions into geospatial coordinates using semantic meshes. This method$\unicode{x2013}$released as a user-friendly open-source toolkit$\unicode{x2013}$enables analysts to use the highest quality data for predictions, capture information about the sides of objects, and leverage multiple viewpoints of each location for added robustness. We demonstrate the value of this approach on a new benchmark dataset of four forest sites in the western U.S. that consists of drone images, photogrammetry results, predicted tree locations, and species classification data derived from manual surveys. We show that our proposed multiview method improves classification accuracy from 53% to 75% relative to an orthomosaic baseline on a challenging cross-site tree species classification task.

The primary color profile of the same identity is assumed to remain consistent in typical Person Re-identification (Person ReID) tasks. However, this assumption may be invalid in real-world situations and images hold variant color profiles, because of cross-modality cameras or identity with different clothing. To address this issue, we propose Color Space Learning (CSL) for those Cross-Color Person ReID problems. Specifically, CSL guides the model to be less color-sensitive with two modules: Image-level Color-Augmentation and Pixel-level Color-Transformation. The first module increases the color diversity of the inputs and guides the model to focus more on the non-color information. The second module projects every pixel of input images onto a new color space. In addition, we introduce a new Person ReID benchmark across RGB and Infrared modalities, NTU-Corridor, which is the first with privacy agreements from all participants. To evaluate the effectiveness and robustness of our proposed CSL, we evaluate it on several Cross-Color Person ReID benchmarks. Our method surpasses the state-of-the-art methods consistently. The code and benchmark are available at: https://github.com/niejiahao1998/CSL

With the advent of image super-resolution (SR) algorithms, how to evaluate the quality of generated SR images has become an urgent task. Although full-reference methods perform well in SR image quality assessment (SR-IQA), their reliance on high-resolution (HR) images limits their practical applicability. Leveraging available reconstruction information as much as possible for SR-IQA, such as low-resolution (LR) images and the scale factors, is a promising way to enhance assessment performance for SR-IQA without HR for reference. In this letter, we attempt to evaluate the perceptual quality and reconstruction fidelity of SR images considering LR images and scale factors. Specifically, we propose a novel dual-branch reduced-reference SR-IQA network, \ie, Perception- and Fidelity-aware SR-IQA (PFIQA). The perception-aware branch evaluates the perceptual quality of SR images by leveraging the merits of global modeling of Vision Transformer (ViT) and local relation of ResNet, and incorporating the scale factor to enable comprehensive visual perception. Meanwhile, the fidelity-aware branch assesses the reconstruction fidelity between LR and SR images through their visual perception. The combination of the two branches substantially aligns with the human visual system, enabling a comprehensive SR image evaluation. Experimental results indicate that our PFIQA outperforms current state-of-the-art models across three widely-used SR-IQA benchmarks. Notably, PFIQA excels in assessing the quality of real-world SR images.

Accurate detection of vulvovaginal candidiasis is critical for women's health, yet its sparse distribution and visually ambiguous characteristics pose significant challenges for accurate identification by pathologists and neural networks alike. Our eye-tracking data reveals that areas garnering sustained attention - yet not marked by experts after deliberation - are often aligned with false positives of neural networks. Leveraging this finding, we introduce Gaze-DETR, a pioneering method that integrates gaze data to enhance neural network precision by diminishing false positives. Gaze-DETR incorporates a universal gaze-guided warm-up protocol applicable across various detection methods and a gaze-guided rectification strategy specifically designed for DETR-based models. Our comprehensive tests confirm that Gaze-DETR surpasses existing leading methods, showcasing remarkable improvements in detection accuracy and generalizability.

Glass largely blurs the boundary between the real world and the reflection. The special transmittance and reflectance quality have confused the semantic tasks related to machine vision. Therefore, how to clear the boundary built by glass, and avoid over-capturing features as false positive information in deep structure, matters for constraining the segmentation of reflection surface and penetrating glass. We proposed the Fourier Boundary Features Network with Wider Catchers (FBWC), which might be the first attempt to utilize sufficiently wide horizontal shallow branches without vertical deepening for guiding the fine granularity segmentation boundary through primary glass semantic information. Specifically, we designed the Wider Coarse-Catchers (WCC) for anchoring large area segmentation and reducing excessive extraction from a structural perspective. We embed fine-grained features by Cross Transpose Attention (CTA), which is introduced to avoid the incomplete area within the boundary caused by reflection noise. For excavating glass features and balancing high-low layers context, a learnable Fourier Convolution Controller (FCC) is proposed to regulate information integration robustly. The proposed method has been validated on three different public glass segmentation datasets. Experimental results reveal that the proposed method yields better segmentation performance compared with the state-of-the-art (SOTA) methods in glass image segmentation.

3D content creation plays a vital role in various applications, such as gaming, robotics simulation, and virtual reality. However, the process is labor-intensive and time-consuming, requiring skilled designers to invest considerable effort in creating a single 3D asset. To address this challenge, text-to-3D generation technologies have emerged as a promising solution for automating 3D creation. Leveraging the success of large vision language models, these techniques aim to generate 3D content based on textual descriptions. Despite recent advancements in this area, existing solutions still face significant limitations in terms of generation quality and efficiency. In this survey, we conduct an in-depth investigation of the latest text-to-3D creation methods. We provide a comprehensive background on text-to-3D creation, including discussions on datasets employed in training and evaluation metrics used to assess the quality of generated 3D models. Then, we delve into the various 3D representations that serve as the foundation for the 3D generation process. Furthermore, we present a thorough comparison of the rapidly growing literature on generative pipelines, categorizing them into feedforward generators, optimization-based generation, and view reconstruction approaches. By examining the strengths and weaknesses of these methods, we aim to shed light on their respective capabilities and limitations. Lastly, we point out several promising avenues for future research. With this survey, we hope to inspire researchers further to explore the potential of open-vocabulary text-conditioned 3D content creation.

Objective: Federated Learning (FL) enables collaborative model training while keeping data locally. Currently, most FL studies in radiology are conducted in simulated environments due to numerous hurdles impeding its translation into practice. The few existing real-world FL initiatives rarely communicate specific measures taken to overcome these hurdles, leaving behind a significant knowledge gap. Minding efforts to implement real-world FL, there is a notable lack of comprehensive assessment comparing FL to less complex alternatives. Materials & Methods: We extensively reviewed FL literature, categorizing insights along with our findings according to their nature and phase while establishing a FL initiative, summarized to a comprehensive guide. We developed our own FL infrastructure within the German Radiological Cooperative Network (RACOON) and demonstrated its functionality by training FL models on lung pathology segmentation tasks across six university hospitals. We extensively evaluated FL against less complex alternatives in three distinct evaluation scenarios. Results: The proposed guide outlines essential steps, identified hurdles, and proposed solutions for establishing successful FL initiatives conducting real-world experiments. Our experimental results show that FL outperforms less complex alternatives in all evaluation scenarios, justifying the effort required to translate FL into real-world applications. Discussion & Conclusion: Our proposed guide aims to aid future FL researchers in circumventing pitfalls and accelerating translation of FL into radiological applications. Our results underscore the value of efforts needed to translate FL into real-world applications by demonstrating advantageous performance over alternatives, and emphasize the importance of strategic organization, robust management of distributed data and infrastructure in real-world settings.

Contrastive pretraining provides robust representations by ensuring their invariance to different image transformations while simultaneously preventing representational collapse. Equivariant contrastive learning, on the other hand, provides representations sensitive to specific image transformations while remaining invariant to others. By introducing equivariance to time-induced transformations, such as disease-related anatomical changes in longitudinal imaging, the model can effectively capture such changes in the representation space. In this work, we pro-pose a Time-equivariant Contrastive Learning (TC) method. First, an encoder embeds two unlabeled scans from different time points of the same patient into the representation space. Next, a temporal equivariance module is trained to predict the representation of a later visit based on the representation from one of the previous visits and the corresponding time interval with a novel regularization loss term while preserving the invariance property to irrelevant image transformations. On a large longitudinal dataset, our model clearly outperforms existing equivariant contrastive methods in predicting progression from intermediate age-related macular degeneration (AMD) to advanced wet-AMD within a specified time-window.

A fundamental tenet of pattern recognition is that overlap between training and testing sets causes an optimistic accuracy estimate. Deep CNNs for face recognition are trained for N-way classification of the identities in the training set. Accuracy is commonly estimated as average 10-fold classification accuracy on image pairs from test sets such as LFW, CALFW, CPLFW, CFP-FP and AgeDB-30. Because train and test sets have been independently assembled, images and identities in any given test set may also be present in any given training set. In particular, our experiments reveal a surprising degree of identity and image overlap between the LFW family of test sets and the MS1MV2 training set. Our experiments also reveal identity label noise in MS1MV2. We compare accuracy achieved with same-size MS1MV2 subsets that are identity-disjoint and not identity-disjoint with LFW, to reveal the size of the optimistic bias. Using more challenging test sets from the LFW family, we find that the size of the optimistic bias is larger for more challenging test sets. Our results highlight the lack of and the need for identity-disjoint train and test methodology in face recognition research.

This study introduces a groundbreaking optical coherence tomography (OCT) imaging system dedicated for high-throughput screening applications using ex vivo tissue culture. Leveraging OCT's non-invasive, high-resolution capabilities, the system is equipped with a custom-designed motorized platform and tissue detection ability for automated, successive imaging across samples. Transformer-based deep learning segmentation algorithms further ensure robust, consistent, and efficient readouts meeting the standards for screening assays. Validated using retinal explant cultures from a mouse model of retinal degeneration, the system provides robust, rapid, reliable, unbiased, and comprehensive readouts of tissue response to treatments. This fully automated OCT-based system marks a significant advancement in tissue screening, promising to transform drug discovery, as well as other relevant research fields.

Synthetic aperture radar (SAR) is essential in actively acquiring information for Earth observation. SAR Automatic Target Recognition (ATR) focuses on detecting and classifying various target categories under different image conditions. The current deep learning-based SAR ATR methods are typically designed for specific datasets and applications. Various target characteristics, scene background information, and sensor parameters across ATR datasets challenge the generalization of those methods. This paper aims to achieve general SAR ATR based on a foundation model with Self-Supervised Learning (SSL). Our motivation is to break through the specific dataset and condition limitations and obtain universal perceptual capabilities across the target, scene, and sensor. A foundation model named SARATR-X is proposed with the following four aspects: pre-training dataset, model backbone, SSL, and evaluation task. First, we integrated 14 datasets with various target categories and imaging conditions as a pre-training dataset. Second, different model backbones were discussed to find the most suitable approaches for remote-sensing images. Third, we applied two-stage training and SAR gradient features to ensure the diversity and scalability of SARATR-X. Finally, SARATR-X has achieved competitive and superior performance on 5 datasets with 8 task settings, which shows that the foundation model can achieve universal SAR ATR. We believe it is time to embrace fundamental models for SAR image interpretation in the era of increasing big data.

In this paper, we propose a predictive quantifier to estimate the retraining cost of a trained model in distribution shifts. The proposed Aggregated Representation Measure (ARM) quantifies the change in the model's representation from the old to new data distribution. It provides, before actually retraining the model, a single concise index of resources - epochs, energy, and carbon emissions - required for the retraining. This enables reuse of a model with a much lower cost than training a new model from scratch. The experimental results indicate that ARM reasonably predicts retraining costs for varying noise intensities and enables comparisons among multiple model architectures to determine the most cost-effective and sustainable option.

Localizing oneself during endoscopic procedures can be problematic due to the lack of distinguishable textures and landmarks, as well as difficulties due to the endoscopic device such as a limited field of view and challenging lighting conditions. Expert knowledge shaped by years of experience is required for localization within the human body during endoscopic procedures. In this work, we present a deep learning method based on anatomy recognition, that constructs a surgical path in an unsupervised manner from surgical videos, modelling relative location and variations due to different viewing angles. At inference time, the model can map an unseen video's frames on the path and estimate the viewing angle, aiming to provide guidance, for instance, to reach a particular destination. We test the method on a dataset consisting of surgical videos of transsphenoidal adenomectomies, as well as on a synthetic dataset. An online tool that lets researchers upload their surgical videos to obtain anatomy detections and the weights of the trained YOLOv7 model are available at: https://surgicalvision.bmic.ethz.ch.

The multi-scale receptive field and large kernel attention (LKA) module have been shown to significantly improve performance in the lightweight image super-resolution task. However, existing lightweight super-resolution (SR) methods seldom pay attention to designing efficient building block with multi-scale receptive field for local modeling, and their LKA modules face a quadratic increase in computational and memory footprints as the convolutional kernel size increases. To address the first issue, we propose the multi-scale blueprint separable convolutions (MBSConv) as highly efficient building block with multi-scale receptive field, it can focus on the learning for the multi-scale information which is a vital component of discriminative representation. As for the second issue, we revisit the key properties of LKA in which we find that the adjacent direct interaction of local information and long-distance dependencies is crucial to provide remarkable performance. Thus, taking this into account and in order to mitigate the complexity of LKA, we propose a large coordinate kernel attention (LCKA) module which decomposes the 2D convolutional kernels of the depth-wise convolutional layers in LKA into horizontal and vertical 1-D kernels. LCKA enables the adjacent direct interaction of local information and long-distance dependencies not only in the horizontal direction but also in the vertical. Besides, LCKA allows for the direct use of extremely large kernels in the depth-wise convolutional layers to capture more contextual information, which helps to significantly improve the reconstruction performance, and it incurs lower computational complexity and memory footprints. Integrating MBSConv and LCKA, we propose a large coordinate kernel attention network (LCAN).

Image-guided depth completion aims at generating a dense depth map from sparse LiDAR data and RGB image. Recent methods have shown promising performance by reformulating it as a classification problem with two sub-tasks: depth discretization and probability prediction. They divide the depth range into several discrete depth values as depth categories, serving as priors for scene depth distributions. However, previous depth discretization methods are easy to be impacted by depth distribution variations across different scenes, resulting in suboptimal scene depth distribution priors. To address the above problem, we propose a progressive depth decoupling and modulating network, which incrementally decouples the depth range into bins and adaptively generates multi-scale dense depth maps in multiple stages. Specifically, we first design a Bins Initializing Module (BIM) to construct the seed bins by exploring the depth distribution information within a sparse depth map, adapting variations of depth distribution. Then, we devise an incremental depth decoupling branch to progressively refine the depth distribution information from global to local. Meanwhile, an adaptive depth modulating branch is developed to progressively improve the probability representation from coarse-grained to fine-grained. And the bi-directional information interactions are proposed to strengthen the information interaction between those two branches (sub-tasks) for promoting information complementation in each branch. Further, we introduce a multi-scale supervision mechanism to learn the depth distribution information in latent features and enhance the adaptation capability across different scenes. Experimental results on public datasets demonstrate that our method outperforms the state-of-the-art methods. The code will be open-sourced at [this https URL](https://github.com/Cisse-away/PDDM).

While content-based image retrieval (CBIR) has been extensively studied in natural image retrieval, its application to medical images presents ongoing challenges, primarily due to the 3D nature of medical images. Recent studies have shown the potential use of pre-trained vision embeddings for CBIR in the context of radiology image retrieval. However, a benchmark for the retrieval of 3D volumetric medical images is still lacking, hindering the ability to objectively evaluate and compare the efficiency of proposed CBIR approaches in medical imaging. In this study, we extend previous work and establish a benchmark for region-based and multi-organ retrieval using the TotalSegmentator dataset (TS) with detailed multi-organ annotations. We benchmark embeddings derived from pre-trained supervised models on medical images against embeddings derived from pre-trained unsupervised models on non-medical images for 29 coarse and 104 detailed anatomical structures in volume and region levels. We adopt a late interaction re-ranking method inspired by text matching for image retrieval, comparing it against the original method proposed for volume and region retrieval achieving retrieval recall of 1.0 for diverse anatomical regions with a wide size range. The findings and methodologies presented in this paper provide essential insights and benchmarks for the development and evaluation of CBIR approaches in the context of medical imaging.

Recent advances in computed tomography (CT) imaging, especially with dual-robot systems, have introduced new challenges for scan trajectory optimization. This paper presents a novel approach using Gated Recurrent Units (GRUs) to optimize CT scan trajectories. Our approach exploits the flexibility of robotic CT systems to select projections that enhance image quality by improving resolution and contrast while reducing scan time. We focus on cone-beam CT and employ several projection-based metrics, including absorption, pixel intensities, contrast-to-noise ratio, and data completeness. The GRU network aims to minimize data redundancy and maximize completeness with a limited number of projections. We validate our method using simulated data of a test specimen, focusing on a specific voxel of interest. The results show that the GRU-optimized scan trajectories can outperform traditional circular CT trajectories in terms of image quality metrics. For the used specimen, SSIM improves from 0.38 to 0.49 and CNR increases from 6.97 to 9.08. This finding suggests that the application of GRU in CT scan trajectory optimization can lead to more efficient, cost-effective, and high-quality imaging solutions.

This paper explores a novel multi-modal alternating learning paradigm pursuing a reconciliation between the exploitation of uni-modal features and the exploration of cross-modal interactions. This is motivated by the fact that current paradigms of multi-modal learning tend to explore multi-modal features simultaneously. The resulting gradient prohibits further exploitation of the features in the weak modality, leading to modality competition, where the dominant modality overpowers the learning process. To address this issue, we study the modality-alternating learning paradigm to achieve reconcilement. Specifically, we propose a new method called ReconBoost to update a fixed modality each time. Herein, the learning objective is dynamically adjusted with a reconcilement regularization against competition with the historical models. By choosing a KL-based reconcilement, we show that the proposed method resembles Friedman's Gradient-Boosting (GB) algorithm, where the updated learner can correct errors made by others and help enhance the overall performance. The major difference with the classic GB is that we only preserve the newest model for each modality to avoid overfitting caused by ensembling strong learners. Furthermore, we propose a memory consolidation scheme and a global rectification scheme to make this strategy more effective. Experiments over six multi-modal benchmarks speak to the efficacy of the method. We release the code at https://github.com/huacong/ReconBoost.

AI-based analysis of histopathology whole slide images (WSIs) is central in computational pathology. However, image quality can impact model performance. Here, we investigate to what extent unsharp areas of WSIs impact deep convolutional neural network classification performance. We propose a multi-model approach, i.e. DeepBlurMM, to alleviate the impact of unsharp image areas and improve the model performance. DeepBlurMM uses the sigma cut-offs to determine the most suitable model for predicting tiles with various levels of blurring within a single WSI, where sigma is the standard deviation of the Gaussian distribution. Specifically, the cut-offs categorise the tiles into sharp or slight blur, moderate blur, and high blur. Each blur level has a corresponding model to be selected for tile-level predictions. Throughout the simulation study, we demonstrated the application of DeepBlurMM in a binary classification task for breast cancer Nottingham Histological Grade 1 vs 3. Performance, evaluated over 5-fold cross-validation, showed that DeepBlurMM outperformed the base model under moderate blur and mixed blur conditions. Unsharp image tiles (local blurriness) at prediction time reduced model performance. The proposed multi-model approach improved performance under some conditions, with the potential to improve quality in both research and clinical applications.

With the benefit of deep learning techniques, recent researches have made significant progress in image compression artifacts reduction. Despite their improved performances, prevailing methods only focus on learning a mapping from the compressed image to the original one but ignore the intrinsic attributes of the given compressed images, which greatly harms the performance of downstream parsing tasks. Different from these methods, we propose to decouple the intrinsic attributes into two complementary features for artifacts reduction,ie, the compression-insensitive features to regularize the high-level semantic representations during training and the compression-sensitive features to be aware of the compression degree. To achieve this, we first employ adversarial training to regularize the compressed and original encoded features for retaining high-level semantics, and we then develop the compression quality-aware feature encoder for compression-sensitive features. Based on these dual complementary features, we propose a Dual Awareness Guidance Network (DAGN) to utilize these awareness features as transformation guidance during the decoding phase. In our proposed DAGN, we develop a cross-feature fusion module to maintain the consistency of compression-insensitive features by fusing compression-insensitive features into the artifacts reduction baseline. Our method achieves an average 2.06 dB PSNR gains on BSD500, outperforming state-of-the-art methods, and only requires 29.7 ms to process one image on BSD500. Besides, the experimental results on LIVE1 and LIU4K also demonstrate the efficiency, effectiveness, and superiority of the proposed method in terms of quantitative metrics, visual quality, and downstream machine vision tasks.

Medical image interpretation using deep learning has shown promise but often requires extensive expert-annotated datasets. To reduce this annotation burden, we develop an Image-Graph Contrastive Learning framework that pairs chest X-rays with structured report knowledge graphs automatically extracted from radiology notes. Our approach uniquely encodes the disconnected graph components via a relational graph convolution network and transformer attention. In experiments on the CheXpert dataset, this novel graph encoding strategy enabled the framework to outperform existing methods that use image-text contrastive learning in 1% linear evaluation and few-shot settings, while achieving comparable performance to radiologists. By exploiting unlabeled paired images and text, our framework demonstrates the potential of structured clinical insights to enhance contrastive learning for medical images. This work points toward reducing demands on medical experts for annotations, improving diagnostic precision, and advancing patient care through robust medical image understanding.

Deep learning classifiers are prone to latching onto dominant confounders present in a dataset rather than on the causal markers associated with the target class, leading to poor generalization and biased predictions. Although explainability via counterfactual image generation has been successful at exposing the problem, bias mitigation strategies that permit accurate explainability in the presence of dominant and diverse artifacts remain unsolved. In this work, we propose the DeCoDEx framework and show how an external, pre-trained binary artifact detector can be leveraged during inference to guide a diffusion-based counterfactual image generator towards accurate explainability. Experiments on the CheXpert dataset, using both synthetic artifacts and real visual artifacts (support devices), show that the proposed method successfully synthesizes the counterfactual images that change the causal pathology markers associated with Pleural Effusion while preserving or ignoring the visual artifacts. Augmentation of ERM and Group-DRO classifiers with the DeCoDEx generated images substantially improves the results across underrepresented groups that are out of distribution for each class. The code is made publicly available at https://github.com/NimaFathi/DeCoDEx.

Recent years have witnessed the rapid development of short videos, which usually contain both visual and audio modalities. Background music is important to the short videos, which can significantly influence the emotions of the viewers. However, at present, the background music of short videos is generally chosen by the video producer, and there is a lack of automatic music recommendation methods for short videos. This paper introduces MVBind, an innovative Music-Video embedding space Binding model for cross-modal retrieval. MVBind operates as a self-supervised approach, acquiring inherent knowledge of intermodal relationships directly from data, without the need of manual annotations. Additionally, to compensate the lack of a corresponding musical-visual pair dataset for short videos, we construct a dataset, SVM-10K(Short Video with Music-10K), which mainly consists of meticulously selected short videos. On this dataset, MVBind manifests significantly improved performance compared to other baseline methods. The constructed dataset and code will be released to facilitate future research.

The task of generating dance from music is crucial, yet current methods, which mainly produce joint sequences, lead to outputs that lack intuitiveness and complicate data collection due to the necessity for precise joint annotations. We introduce a Dance Any Beat Diffusion model, namely DabFusion, that employs music as a conditional input to directly create dance videos from still images, utilizing conditional image-to-video generation principles. This approach pioneers the use of music as a conditioning factor in image-to-video synthesis. Our method unfolds in two stages: training an auto-encoder to predict latent optical flow between reference and driving frames, eliminating the need for joint annotation, and training a U-Net-based diffusion model to produce these latent optical flows guided by music rhythm encoded by CLAP. Although capable of producing high-quality dance videos, the baseline model struggles with rhythm alignment. We enhance the model by adding beat information, improving synchronization. We introduce a 2D motion-music alignment score (2D-MM Align) for quantitative assessment. Evaluated on the AIST++ dataset, our enhanced model shows marked improvements in 2D-MM Align score and established metrics. Video results can be found on our project page: https://DabFusion.github.io.

Image-to-image translation can create large impact in medical imaging, i.e. if images of a patient can be translated to another modality, type or sequence for better diagnosis. However, these methods must be validated by human reader studies, which are costly and restricted to small samples. Automatic evaluation of large samples to pre-evaluate and continuously improve methods before human validation is needed. In this study, we give an overview of reference and non-reference metrics for image synthesis assessment and investigate the ability of nine metrics, that need a reference (SSIM, MS-SSIM, PSNR, MSE, NMSE, MAE, LPIPS, NMI and PCC) and three non-reference metrics (BLUR, MSN, MNG) to detect 11 kinds of distortions in MR images from the BraSyn dataset. In addition we test a downstream segmentation metric and the effect of three normalization methods (Minmax, cMinMax and Zscore). Although PSNR and SSIM are frequently used to evaluate generative models for image-to-image-translation tasks in the medical domain, they show very specific shortcomings. SSIM ignores blurring but is very sensitive to intensity shifts in unnormalized MR images. PSNR is even more sensitive to different normalization methods and hardly measures the degree of distortions. Further metrics, such as LPIPS, NMI and DICE can be very useful to evaluate other similarity aspects. If the images to be compared are misaligned, most metrics are flawed. By carefully selecting and reasonably combining image similarity metrics, the training and selection of generative models for MR image synthesis can be improved. Many aspects of their output can be validated before final and costly evaluation by trained radiologists is conducted.

The graph neural networks has been proved to be an efficient machine learning technique in real life applications. The handwritten recognition is one of the useful area in real life use where both offline and online handwriting recognition are required. The chain code as feature extraction technique has shown significant results in literature and we have been able to use chain codes with graph neural networks. To the best of our knowledge, this work presents first time a novel combination of handwritten trajectories features as chain codes and graph neural networks together. The handwritten trajectories for offline handwritten text has been evaluated using recovery of drawing order, whereas online handwritten trajectories are directly used with chain codes. Our results prove that present combination surpass previous results and minimize error rate in few epochs only.

The rapid advancement of foundation models (FMs) across language, image, audio, and video domains has shown remarkable capabilities in diverse tasks. However, the proliferation of FMs brings forth a critical challenge: the potential to generate hallucinated outputs, particularly in high-stakes applications. The tendency of foundation models to produce hallucinated content arguably represents the biggest hindrance to their widespread adoption in real-world scenarios, especially in domains where reliability and accuracy are paramount. This survey paper presents a comprehensive overview of recent developments that aim to identify and mitigate the problem of hallucination in FMs, spanning text, image, video, and audio modalities. By synthesizing recent advancements in detecting and mitigating hallucination across various modalities, the paper aims to provide valuable insights for researchers, developers, and practitioners. Essentially, it establishes a clear framework encompassing definition, taxonomy, and detection strategies for addressing hallucination in multimodal foundation models, laying the foundation for future research in this pivotal area.

We introduce Xmodel-VLM, a cutting-edge multimodal vision language model. It is designed for efficient deployment on consumer GPU servers. Our work directly confronts a pivotal industry issue by grappling with the prohibitive service costs that hinder the broad adoption of large-scale multimodal systems. Through rigorous training, we have developed a 1B-scale language model from the ground up, employing the LLaVA paradigm for modal alignment. The result, which we call Xmodel-VLM, is a lightweight yet powerful multimodal vision language model. Extensive testing across numerous classic multimodal benchmarks has revealed that despite its smaller size and faster execution, Xmodel-VLM delivers performance comparable to that of larger models. Our model checkpoints and code are publicly available on GitHub at https://github.com/XiaoduoAILab/XmodelVLM.

With recent advances in deep learning, numerous algorithms have been developed to enhance video quality, reduce visual artefacts and improve perceptual quality. However, little research has been reported on the quality assessment of enhanced content - the evaluation of enhancement methods is often based on quality metrics that were designed for compression applications. In this paper, we propose a novel blind deep video quality assessment (VQA) method specifically for enhanced video content. It employs a new Recurrent Memory Transformer (RMT) based network architecture to obtain video quality representations, which is optimised through a novel content-quality-aware contrastive learning strategy based on a new database containing 13K training patches with enhanced content. The extracted quality representations are then combined through linear regression to generate video-level quality indices. The proposed method, RMT-BVQA, has been evaluated on the VDPVE (VQA Dataset for Perceptual Video Enhancement) database through a five-fold cross validation. The results show its superior correlation performance when compared to ten existing no-reference quality metrics.

In this work, we propose to tackle several challenges hindering the development of Automatic Target Detection (ATD) algorithms for ground targets in SAR images. To address the lack of representative training data, we propose a Deep Learning approach to train ATD models with synthetic target signatures produced with the MOCEM simulator. We define an incrustation pipeline to incorporate synthetic targets into real backgrounds. Using this hybrid dataset, we train ATD models specifically tailored to bridge the domain gap between synthetic and real data. Our approach notably relies on massive physics-based data augmentation techniques and Adversarial Training of two deep-learning detection architectures. We then test these models on several datasets, including (1) patchworks of real SAR images, (2) images with the incrustation of real targets in real backgrounds, and (3) images with the incrustation of synthetic background objects in real backgrounds. Results show that the produced hybrid datasets are exempt from image overlay bias. Our approach can reach up to 90% of Average Precision on real data while exclusively using synthetic targets for training.

Due to the increasing need for effective security measures and the integration of cameras in commercial products, a hugeamount of visual data is created today. Law enforcement agencies (LEAs) are inspecting images and videos to findradicalization, propaganda for terrorist organizations and illegal products on darknet markets. This is time consuming.Instead of an undirected search, LEAs would like to adapt to new crimes and threats, and focus only on data from specificlocations, persons or objects, which requires flexible interpretation of image content. Visual concept detection with deepconvolutional neural networks (CNNs) is a crucial component to understand the image content. This paper has fivecontributions. The first contribution allows image-based geo-localization to estimate the origin of an image. CNNs andgeotagged images are used to create a model that determines the location of an image by its pixel values. The secondcontribution enables analysis of fine-grained concepts to distinguish sub-categories in a generic concept. The proposedmethod encompasses data acquisition and cleaning and concept hierarchies. The third contribution is the recognition ofperson attributes (e.g., glasses or moustache) to enable query by textual description for a person. The person-attributeproblem is treated as a specific sub-task of concept classification. The fourth contribution is an intuitive image annotationtool based on active learning. Active learning allows users to define novel concepts flexibly and train CNNs with minimalannotation effort. The fifth contribution increases the flexibility for LEAs in the query definition by using query expansion.Query expansion maps user queries to known and detectable concepts. Therefore, no prior knowledge of the detectableconcepts is required for the users. The methods are validated on data with varying locations (popular and non-touristiclocations), varying person attributes (CelebA dataset), and varying number of annotations.

From a perspective of feature matching, optical flow estimation for event cameras involves identifying event correspondences by comparing feature similarity across accompanying event frames. In this work, we introduces an effective and robust high-dimensional (HD) feature descriptor for event frames, utilizing Vector Symbolic Architectures (VSA). The topological similarity among neighboring variables within VSA contributes to the enhanced representation similarity of feature descriptors for flow-matching points, while its structured symbolic representation capacity facilitates feature fusion from both event polarities and multiple spatial scales. Based on this HD feature descriptor, we propose a novel feature matching framework for event-based optical flow, encompassing both model-based (VSA-Flow) and self-supervised learning (VSA-SM) methods. In VSA-Flow, accurate optical flow estimation validates the effectiveness of HD feature descriptors. In VSA-SM, a novel similarity maximization method based on the HD feature descriptor is proposed to learn optical flow in a self-supervised way from events alone, eliminating the need for auxiliary grayscale images. Evaluation results demonstrate that our VSA-based method achieves superior accuracy in comparison to both model-based and self-supervised learning methods on the DSEC benchmark, while remains competitive among both methods on the MVSEC benchmark. This contribution marks a significant advancement in event-based optical flow within the feature matching methodology.

The automation of writing imaging reports is a valuable tool for alleviating the workload of radiologists. Crucial steps in this process involve the cross-modal alignment between medical images and reports, as well as the retrieval of similar historical cases. However, the presence of presentation-style vocabulary (e.g., sentence structure and grammar) in reports poses challenges for cross-modal alignment. Additionally, existing methods for similar historical cases retrieval face suboptimal performance owing to the modal gap issue. In response, this paper introduces a novel method, named Factual Serialization Enhancement (FSE), for chest X-ray report generation. FSE begins with the structural entities approach to eliminate presentation-style vocabulary in reports, providing specific input for our model. Then, uni-modal features are learned through cross-modal alignment between images and factual serialization in reports. Subsequently, we present a novel approach to retrieve similar historical cases from the training set, leveraging aligned image features. These features implicitly preserve semantic similarity with their corresponding reference reports, enabling us to calculate similarity solely among aligned features. This effectively eliminates the modal gap issue for knowledge retrieval without the requirement for disease labels. Finally, the cross-modal fusion network is employed to query valuable information from these cases, enriching image features and aiding the text decoder in generating high-quality reports. Experiments on MIMIC-CXR and IU X-ray datasets from both specific and general scenarios demonstrate the superiority of FSE over state-of-the-art approaches in both natural language generation and clinical efficacy metrics.

As image recognition models become more prevalent, scalable coding methods for machines and humans gain more importance. Applications of image recognition models include traffic monitoring and farm management. In these use cases, the scalable coding method proves effective because the tasks require occasional image checking by humans. Existing image compression methods for humans and machines meet these requirements to some extent. However, these compression methods are effective solely for specific image recognition models. We propose a learning-based scalable image coding method for humans and machines that is compatible with numerous image recognition models. We combine an image compression model for machines with a compression model, providing additional information to facilitate image decoding for humans. The features in these compression models are fused using a feature fusion network to achieve efficient image compression. Our method's additional information compression model is adjusted to reduce the number of parameters by enabling combinations of features of different sizes in the feature fusion network. Our approach confirms that the feature fusion network efficiently combines image compression models while reducing the number of parameters. Furthermore, we demonstrate the effectiveness of the proposed scalable coding method by evaluating the image compression performance in terms of decoded image quality and bitrate.

Most dataset distillation methods struggle to accommodate large-scale datasets due to their substantial computational and memory requirements. In this paper, we present a curriculum-based dataset distillation framework designed to harmonize scalability with efficiency. This framework strategically distills synthetic images, adhering to a curriculum that transitions from simple to complex. By incorporating curriculum evaluation, we address the issue of previous methods generating images that tend to be homogeneous and simplistic, doing so at a manageable computational cost. Furthermore, we introduce adversarial optimization towards synthetic images to further improve their representativeness and safeguard against their overfitting to the neural network involved in distilling. This enhances the generalization capability of the distilled images across various neural network architectures and also increases their robustness to noise. Extensive experiments demonstrate that our framework sets new benchmarks in large-scale dataset distillation, achieving substantial improvements of 11.1\% on Tiny-ImageNet, 9.0\% on ImageNet-1K, and 7.3\% on ImageNet-21K. The source code will be released to the community.

Anomaly detection and localization without any manual annotations and prior knowledge is a challenging task under the setting of unsupervised learning. The existing works achieve excellent performance in the anomaly detection, but with complex networks or cumbersome pipelines. To address this issue, this paper explores a simple but effective architecture in the anomaly detection. It consists of a well pre-trained encoder to extract hierarchical feature representations and a decoder to reconstruct these intermediate features from the encoder. In particular, it does not require any data augmentations and anomalous images for training. The anomalies can be detected when the decoder fails to reconstruct features well, and then errors of hierarchical feature reconstruction are aggregated into an anomaly map to achieve anomaly localization. The difference comparison between those features of encoder and decode lead to more accurate and robust localization results than the comparison in single feature or pixel-by-pixel comparison in the conventional works. Experiment results show that the proposed method outperforms the state-of-the-art methods on MNIST, Fashion-MNIST, CIFAR-10, and MVTec Anomaly Detection datasets on both anomaly detection and localization.

Gait recognition, a rapidly advancing vision technology for person identification from a distance, has made significant strides in indoor settings. However, evidence suggests that existing methods often yield unsatisfactory results when applied to newly released real-world gait datasets. Furthermore, conclusions drawn from indoor gait datasets may not easily generalize to outdoor ones. Therefore, the primary goal of this work is to present a comprehensive benchmark study aimed at improving practicality rather than solely focusing on enhancing performance. To this end, we first develop OpenGait, a flexible and efficient gait recognition platform. Using OpenGait as a foundation, we conduct in-depth ablation experiments to revisit recent developments in gait recognition. Surprisingly, we detect some imperfect parts of certain prior methods thereby resulting in several critical yet undiscovered insights. Inspired by these findings, we develop three structurally simple yet empirically powerful and practically robust baseline models, i.e., DeepGaitV2, SkeletonGait, and SkeletonGait++, respectively representing the appearance-based, model-based, and multi-modal methodology for gait pattern description. Beyond achieving SoTA performances, more importantly, our careful exploration sheds new light on the modeling experience of deep gait models, the representational capacity of typical gait modalities, and so on. We hope this work can inspire further research and application of gait recognition towards better practicality. The code is available at https://github.com/ShiqiYu/OpenGait.

Despite the impressive performance of Multi-view Stereo (MVS) approaches given plenty of training samples, the performance degradation when generalizing to unseen domains has not been clearly explored yet. In this work, we focus on the domain generalization problem in MVS. To evaluate the generalization results, we build a novel MVS domain generalization benchmark including synthetic and real-world datasets. In contrast to conventional domain generalization benchmarks, we consider a more realistic but challenging scenario, where only one source domain is available for training. The MVS problem can be analogized back to the feature matching task, and maintaining robust feature consistency among views is an important factor for improving generalization performance. To address the domain generalization problem in MVS, we propose a novel MVS framework, namely RobustMVS. A DepthClustering-guided Whitening (DCW) loss is further introduced to preserve the feature consistency among different views, which decorrelates multi-view features from viewpoint-specific style information based on geometric priors from depth maps. The experimental results further show that our method achieves superior performance on the domain generalization benchmark.

Internal Language Model (LM)-based methods use permutation language modeling (PLM) to solve the error correction caused by conditional independence in external LM-based methods. However, random permutations of human interference cause fit oscillations in the model training, and Iterative Refinement (IR) operation to improve multimodal information decoupling also introduces additional overhead. To address these issues, this paper proposes the Hierarchical Attention autoregressive Model with Adaptive Permutation (HAAP) to enhance the location-context-image interaction capability, improving autoregressive generalization with internal LM. First, we propose Implicit Permutation Neurons (IPN) to generate adaptive attention masks to dynamically exploit token dependencies. The adaptive masks increase the diversity of training data and prevent model dependency on a specific order. It reduces the training overhead of PLM while avoiding training fit oscillations. Second, we develop Cross-modal Hierarchical Attention mechanism (CHA) to couple context and image features. This processing establishes rich positional semantic dependencies between context and image while avoiding IR. Extensive experimental results show the proposed HAAP achieves state-of-the-art (SOTA) performance in terms of accuracy, complexity, and latency on several datasets.

In this paper, we delve into a new task known as small object editing (SOE), which focuses on text-based image inpainting within a constrained, small-sized area. Despite the remarkable success have been achieved by current image inpainting approaches, their application to the SOE task generally results in failure cases such as Object Missing, Text-Image Mismatch, and Distortion. These failures stem from the limited use of small-sized objects in training datasets and the downsampling operations employed by U-Net models, which hinders accurate generation. To overcome these challenges, we introduce a novel training-based approach, SOEDiff, aimed at enhancing the capability of baseline models like StableDiffusion in editing small-sized objects while minimizing training costs. Specifically, our method involves two key components: SO-LoRA, which efficiently fine-tunes low-rank matrices, and Cross-Scale Score Distillation loss, which leverages high-resolution predictions from the pre-trained teacher diffusion model. Our method presents significant improvements on the test dataset collected from MSCOCO and OpenImage, validating the effectiveness of our proposed method in small object editing. In particular, when comparing SOEDiff with SD-I model on the OpenImage-f dataset, we observe a 0.99 improvement in CLIP-Score and a reduction of 2.87 in FID. Our project page can be found in https://soediff.github.io/.

Haze severely degrades the visual quality of remote sensing images and hampers the performance of automotive navigation, intelligent monitoring, and urban management. The emerging denoising diffusion probabilistic model (DDPM) exhibits the significant potential for dense haze removal with its strong generation ability. Since remote sensing images contain extensive small-scale texture structures, it is important to effectively restore image details from hazy images. However, current wisdom of DDPM fails to preserve image details and color fidelity well, limiting its dehazing capacity for remote sensing images. In this paper, we propose a novel unified Fourier-aware diffusion model for remote sensing image dehazing, termed RSHazeDiff. From a new perspective, RSHazeDiff explores the conditional DDPM to improve image quality in dense hazy scenarios, and it makes three key contributions. First, RSHazeDiff refines the training phase of diffusion process by performing noise estimation and reconstruction constraints in a coarse-to-fine fashion. Thus, it remedies the unpleasing results caused by the simple noise estimation constraint in DDPM. Second, by taking the frequency information as important prior knowledge during iterative sampling steps, RSHazeDiff can preserve more texture details and color fidelity in dehazed images. Third, we design a global compensated learning module to utilize the Fourier transform to capture the global dependency features of input images, which can effectively mitigate the effects of boundary artifacts when processing fixed-size patches. Experiments on both synthetic and real-world benchmarks validate the favorable performance of RSHazeDiff over multiple state-of-the-art methods. Source code will be released at https://github.com/jm-xiong/RSHazeDiff.

Deep models produce a number of features in each internal layer. A key problem in applications such as feature compression for remote inference is determining how important each feature is for the task(s) performed by the model. The problem is especially challenging in the case of multi-task inference, where the same feature may carry different importance for different tasks. In this paper, we examine how effective is mutual information (MI) between a feature and a model's task output as a measure of the feature's importance for that task. Experiments involving hard selection and soft selection (unequal compression) based on MI are carried out to compare the MI-based method with alternative approaches. Multi-objective analysis is provided to offer further insight.

Although face analysis has achieved remarkable improvements in the past few years, designing a multi-task face analysis model is still challenging. Most face analysis tasks are studied as separate problems and do not benefit from the synergy among related tasks. In this work, we propose a novel task-adaptive multi-task face analysis method named as Q-Face, which simultaneously performs multiple face analysis tasks with a unified model. We fuse the features from multiple layers of a large-scale pre-trained model so that the whole model can use both local and global facial information to support multiple tasks. Furthermore, we design a task-adaptive module that performs cross-attention between a set of query vectors and the fused multi-stage features and finally adaptively extracts desired features for each face analysis task. Extensive experiments show that our method can perform multiple tasks simultaneously and achieves state-of-the-art performance on face expression recognition, action unit detection, face attribute analysis, age estimation, and face pose estimation. Compared to conventional methods, our method opens up new possibilities for multi-task face analysis and shows the potential for both accuracy and efficiency.

In medical image segmentation tasks, diffusion models have shown significant potential. However, mainstream diffusion models suffer from drawbacks such as multiple sampling times and slow prediction results. Recently, consistency models, as a standalone generative network, have resolved this issue. Compared to diffusion models, consistency models can reduce the sampling times to once, not only achieving similar generative effects but also significantly speeding up training and prediction. However, they are not suitable for image segmentation tasks, and their application in the medical imaging field has not yet been explored. Therefore, this paper applies the consistency model to medical image segmentation tasks, designing multi-scale feature signal supervision modes and loss function guidance to achieve model convergence. Experiments have verified that the CTS model can obtain better medical image segmentation results with a single sampling during the test phase.

The detection and tracking of small targets in passive optical remote sensing (PORS) has broad applications. However, most of the previously proposed methods seldom utilize the abundant temporal features formed by target motion, resulting in poor detection and tracking performance for low signal-to-clutter ratio (SCR) targets. In this article, we analyze the difficulty based on spatial features and the feasibility based on temporal features of realizing effective detection. According to this analysis, we use a multi-frame as a detection unit and propose a detection method based on temporal energy selective scaling (TESS). Specifically, we investigated the composition of intensity temporal profiles (ITPs) formed by pixels on a multi-frame detection unit. For the target-present pixel, the target passing through the pixel will bring a weak transient disturbance on the ITP and introduce a change in the statistical properties of ITP. We use a well-designed function to amplify the transient disturbance, suppress the background and noise components, and output the trajectory of the target on the multi-frame detection unit. Subsequently, to solve the contradiction between the detection rate and the false alarm rate brought by the traditional threshold segmentation, we associate the temporal and spatial features of the output trajectory and propose a trajectory extraction method based on the 3D Hough transform. Finally, we model the trajectory of the target and propose a trajectory-based multi-target tracking method. Compared with the various state-of-the-art detection and tracking methods, experiments in multiple scenarios prove the superiority of our proposed methods.

Recent advancements in deep learning for 3D models have propelled breakthroughs in generation, detection, and scene understanding. However, the effectiveness of these algorithms hinges on large training datasets. We address the challenge by introducing Efficient 3D Seam Carving (E3SC), a novel 3D model augmentation method based on seam carving, which progressively deforms only part of the input model while ensuring the overall semantics are unchanged. Experiments show that our approach is capable of producing diverse and high-quality augmented 3D shapes across various types and styles of input models, achieving considerable improvements over previous methods. Quantitative evaluations demonstrate that our method effectively enhances the novelty and quality of shapes generated by other subsequent 3D generation algorithms.

This study investigates the use of trajectory and dynamic state information for efficient data curation in autonomous driving machine learning tasks. We propose methods for clustering trajectory-states and sampling strategies in an active learning framework, aiming to reduce annotation and data costs while maintaining model performance. Our approach leverages trajectory information to guide data selection, promoting diversity in the training data. We demonstrate the effectiveness of our methods on the trajectory prediction task using the nuScenes dataset, showing consistent performance gains over random sampling across different data pool sizes, and even reaching sub-baseline displacement errors at just 50% of the data cost. Our results suggest that sampling typical data initially helps overcome the ''cold start problem,'' while introducing novelty becomes more beneficial as the training pool size increases. By integrating trajectory-state-informed active learning, we demonstrate that more efficient and robust autonomous driving systems are possible and practical using low-cost data curation strategies.

Today's analog/mixed-signal (AMS) integrated circuit (IC) designs demand substantial manual intervention. The advent of multimodal large language models (MLLMs) has unveiled significant potential across various fields, suggesting their applicability in streamlining large-scale AMS IC design as well. A bottleneck in employing MLLMs for automatic AMS circuit generation is the absence of a comprehensive dataset delineating the schematic-netlist relationship. We therefore design an automatic technique for converting schematics into netlists, and create dataset AMSNet, encompassing transistor-level schematics and corresponding SPICE format netlists. With a growing size, AMSNet can significantly facilitate exploration of MLLM applications in AMS circuit design. We have made an initial set of netlists public, and will make both our netlist generation tool and the full dataset available upon publishing of this paper.

Domain adaptation is pivotal for enabling deep learning models to generalize across diverse domains, a task complicated by variations in presentation and cognitive nuances. In this paper, we introduce AD-Aligning, a novel approach that combines adversarial training with source-target domain alignment to enhance generalization capabilities. By pretraining with Coral loss and standard loss, AD-Aligning aligns target domain statistics with those of the pretrained encoder, preserving robustness while accommodating domain shifts. Through extensive experiments on diverse datasets and domain shift scenarios, including noise-induced shifts and cognitive domain adaptation tasks, we demonstrate AD-Aligning's superior performance compared to existing methods such as Deep Coral and ADDA. Our findings highlight AD-Aligning's ability to emulate the nuanced cognitive processes inherent in human perception, making it a promising solution for real-world applications requiring adaptable and robust domain adaptation strategies.

In this paper, we address the segmentation of tumor subtypes in whole slide images (WSI) by utilizing incomplete label proportions. Specifically, we utilize `partial' label proportions, which give the proportions among tumor subtypes but do not give the proportion between tumor and non-tumor. Partial label proportions are recorded as the standard diagnostic information by pathologists, and we, therefore, want to use them for realizing the segmentation model that can classify each WSI patch into one of the tumor subtypes or non-tumor. We call this problem ``learning from partial label proportions (LPLP)'' and formulate the problem as a weakly supervised learning problem. Then, we propose an efficient algorithm for this challenging problem by decomposing it into two weakly supervised learning subproblems: multiple instance learning (MIL) and learning from label proportions (LLP). These subproblems are optimized efficiently in the end-to-end manner. The effectiveness of our algorithm is demonstrated through experiments conducted on two WSI datasets.

Significant progress has been made in the field of handwritten mathematical expression recognition, while existing encoder-decoder methods are usually difficult to model global information in \LaTeX. Therefore, this paper introduces a novel approach, Implicit Character-Aided Learning (ICAL), to mine the global expression information and enhance handwritten mathematical expression recognition. Specifically, we propose the Implicit Character Construction Module (ICCM) to predict implicit character sequences and use a Fusion Module to merge the outputs of the ICCM and the decoder, thereby producing corrected predictions. By modeling and utilizing implicit character information, ICAL achieves a more accurate and context-aware interpretation of handwritten mathematical expressions. Experimental results demonstrate that ICAL notably surpasses the state-of-the-art(SOTA) models, improving the expression recognition rate (ExpRate) by 2.21\%/1.75\%/1.28\% on the CROHME 2014/2016/2019 datasets respectively, and achieves a remarkable 69.25\% on the challenging HME100k test set. We make our code available on the GitHub: https://github.com/qingzhenduyu/ICAL

The ambiguous appearance, tiny scale, and fine-grained classes of objects in remote sensing imagery inevitably lead to the noisy annotations in category labels of detection dataset. However, the effects and treatments of the label noises are underexplored in modern oriented remote sensing object detectors. To address this issue, we propose a robust oriented remote sensing object detection method through dynamic loss decay (DLD) mechanism, inspired by the two phase ``early-learning'' and ``memorization'' learning dynamics of deep neural networks on clean and noisy samples. To be specific, we first observe the end point of early learning phase termed as EL, after which the models begin to memorize the false labels that significantly degrade the detection accuracy. Secondly, under the guidance of the training indicator, the losses of each sample are ranked in descending order, and we adaptively decay the losses of the top K largest ones (bad samples) in the following epochs. Because these large losses are of high confidence to be calculated with wrong labels. Experimental results show that the method achieves excellent noise resistance performance tested on multiple public datasets such as HRSC2016 and DOTA-v1.0/v2.0 with synthetic category label noise. Our solution also has won the 2st place in the "fine-grained object detection based on sub-meter remote sensing imagery" track with noisy labels of 2023 National Big Data and Computing Intelligence Challenge.

Referring Image Segmentation (RIS) consistently requires language and appearance semantics to more understand each other. The need becomes acute especially under hard situations. To achieve, existing works tend to resort to various trans-representing mechanisms to directly feed forward language semantic along main RGB branch, which however will result in referent distribution weakly-mined in space and non-referent semantic contaminated along channel. In this paper, we propose Spatial Semantic Recurrent Mining (S\textsuperscript{2}RM) to achieve high-quality cross-modality fusion. It follows a working strategy of trilogy: distributing language feature, spatial semantic recurrent coparsing, and parsed-semantic balancing. During fusion, S\textsuperscript{2}RM will first generate a constraint-weak yet distribution-aware language feature, then bundle features of each row and column from rotated features of one modality context to recurrently correlate relevant semantic contained in feature from other modality context, and finally resort to self-distilled weights to weigh on the contributions of different parsed semantics. Via coparsing, S\textsuperscript{2}RM transports information from the near and remote slice layers of generator context to the current slice layer of parsed context, capable of better modeling global relationship bidirectional and structured. Besides, we also propose a Cross-scale Abstract Semantic Guided Decoder (CASG) to emphasize the foreground of the referent, finally integrating different grained features at a comparatively low cost. Extensive experimental results on four current challenging datasets show that our proposed method performs favorably against other state-of-the-art algorithms.

We investigate the problem of pixelwise correspondence for deformable objects, namely cloth and rope, by comparing both classical and learning-based methods. We choose cloth and rope because they are traditionally some of the most difficult deformable objects to analytically model with their large configuration space, and they are meaningful in the context of robotic tasks like cloth folding, rope knot-tying, T-shirt folding, curtain closing, etc. The correspondence problem is heavily motivated in robotics, with wide-ranging applications including semantic grasping, object tracking, and manipulation policies built on top of correspondences. We present an exhaustive survey of existing classical methods for doing correspondence via feature-matching, including SIFT, SURF, and ORB, and two recently published learning-based methods including TimeCycle and Dense Object Nets. We make three main contributions: (1) a framework for simulating and rendering synthetic images of deformable objects, with qualitative results demonstrating transfer between our simulated and real domains (2) a new learning-based correspondence method extending Dense Object Nets, and (3) a standardized comparison across state-of-the-art correspondence methods. Our proposed method provides a flexible, general formulation for learning temporally and spatially continuous correspondences for nonrigid (and rigid) objects. We report root mean squared error statistics for all methods and find that Dense Object Nets outperforms baseline classical methods for correspondence, and our proposed extension of Dense Object Nets performs similarly.

"How does the person in the bounding box feel?" Achieving human-level recognition of the apparent emotion of a person in real world situations remains an unsolved task in computer vision. Facial expressions are not enough: body pose, contextual knowledge, and commonsense reasoning all contribute to how humans perform this emotional theory of mind task. In this paper, we examine two major approaches enabled by recent large vision language models: 1) image captioning followed by a language-only LLM, and 2) vision language models, under zero-shot and fine-tuned setups. We evaluate the methods on the Emotions in Context (EMOTIC) dataset and demonstrate that a vision language model, fine-tuned even on a small dataset, can significantly outperform traditional baselines. The results of this work aim to help robots and agents perform emotionally sensitive decision-making and interaction in the future.

We perform detailed theoretical analysis of an expectation-maximization-based algorithm recently proposed in for solving a variation of the 3D registration problem, named multi-model 3D registration. Despite having shown superior empirical results, did not theoretically justify the conditions under which the EM approach converges to the ground truth. In this project, we aim to close this gap by establishing such conditions. In particular, the analysis revolves around the usage of probabilistic tail bounds that are developed and applied in various instances throughout the course. The problem studied in this project stands as another example, different from those seen in the course, in which tail-bounds help advance our algorithmic understanding in a probabilistic way. We provide self-contained background materials on 3D Registration

Modeling visual saliency in graphical user interfaces (GUIs) allows to understand how people perceive GUI designs and what elements attract their attention. One aspect that is often overlooked is the fact that computational models depend on a series of design parameters that are not straightforward to decide. We systematically analyze how different design parameters affect scanpath evaluation metrics using a state-of-the-art computational model (DeepGaze++). We particularly focus on three design parameters: input image size, inhibition-of-return decay, and masking radius. We show that even small variations of these design parameters have a noticeable impact on standard evaluation metrics such as DTW or Eyenalysis. These effects also occur in other scanpath models, such as UMSS and ScanGAN, and in other datasets such as MASSVIS. Taken together, our results put forward the impact of design decisions for predicting users' viewing behavior on GUIs.

In recent years, street view imagery has grown to become one of the most important sources of geospatial data collection and urban analytics, which facilitates generating meaningful insights and assisting in decision-making. Synthesizing a street-view image from its corresponding satellite image is a challenging task due to the significant differences in appearance and viewpoint between the two domains. In this study, we screened 20 recent research papers to provide a thorough review of the state-of-the-art of how street-view images are synthesized from their corresponding satellite counterparts. The main findings are: (i) novel deep learning techniques are required for synthesizing more realistic and accurate street-view images; (ii) more datasets need to be collected for public usage; and (iii) more specific evaluation metrics need to be investigated for evaluating the generated images appropriately. We conclude that, due to applying outdated deep learning techniques, the recent literature failed to generate detailed and diverse street-view images.

This paper proposes leveraging vision-language pretraining on bone X-rays paired with French reports to address downstream tasks of interest on bone radiography. A practical processing pipeline is introduced to anonymize and process French medical reports. Pretraining then consists in the self-supervised alignment of visual and textual embedding spaces derived from deep model encoders. The resulting image encoder is then used to handle various downstream tasks, including quantification of osteoarthritis, estimation of bone age on pediatric wrists, bone fracture and anomaly detection. Our approach demonstrates competitive performance on downstream tasks, compared to alternatives requiring a significantly larger amount of human expert annotations. Our work stands as the first study to integrate French reports to shape the embedding space devoted to bone X-Rays representations, capitalizing on the large quantity of paired images and reports data available in an hospital. By relying on generic vision-laguage deep models in a language-specific scenario, it contributes to the deployement of vision models for wider healthcare applications.

CLIP models perform remarkably well on zero-shot classification and retrieval tasks. But recent studies have shown that learnt representations in CLIP are not well suited for dense prediction tasks like object detection, semantic segmentation or depth estimation. More recently, multi-stage training methods for CLIP models was introduced to mitigate the weak performance of CLIP on downstream tasks. In this work, we find that simply improving the quality of captions in image-text datasets improves the quality of CLIP's visual representations, resulting in significant improvement on downstream dense prediction vision tasks. In fact, we find that CLIP pretraining with good quality captions can surpass recent supervised, self-supervised and weakly supervised pretraining methods. We show that when CLIP model with ViT-B/16 as image encoder is trained on well aligned image-text pairs it obtains 12.1% higher mIoU and 11.5% lower RMSE on semantic segmentation and depth estimation tasks over recent state-of-the-art Masked Image Modeling (MIM) pretraining methods like Masked Autoencoder (MAE). We find that mobile architectures also benefit significantly from CLIP pretraining. A recent mobile vision architecture, MCi2, with CLIP pretraining obtains similar performance as Swin-L, pretrained on ImageNet-22k for semantic segmentation task while being 6.1$\times$ smaller. Moreover, we show that improving caption quality results in $10\times$ data efficiency when finetuning for dense prediction tasks.

Many query-based approaches for 3D Multi-Object Tracking (MOT) adopt the tracking-by-attention paradigm, utilizing track queries for identity-consistent detection and object queries for identity-agnostic track spawning. Tracking-by-attention, however, entangles detection and tracking queries in one embedding for both the detection and tracking task, which is sub-optimal. Other approaches resemble the tracking-by-detection paradigm, detecting objects using decoupled track and detection queries followed by a subsequent association. These methods, however, do not leverage synergies between the detection and association task. Combining the strengths of both paradigms, we introduce ADA-Track, a novel end-to-end framework for 3D MOT from multi-view cameras. We introduce a learnable data association module based on edge-augmented cross-attention, leveraging appearance and geometric features. Furthermore, we integrate this association module into the decoder layer of a DETR-based 3D detector, enabling simultaneous DETR-like query-to-image cross-attention for detection and query-to-query cross-attention for data association. By stacking these decoder layers, queries are refined for the detection and association task alternately, effectively harnessing the task dependencies. We evaluate our method on the nuScenes dataset and demonstrate the advantage of our approach compared to the two previous paradigms. Code is available at https://github.com/dsx0511/ADA-Track.

Current video summarization methods primarily depend on supervised computer vision techniques, which demands time-consuming manual annotations. Further, the annotations are always subjective which make this task more challenging. To address these issues, we analyzed the feasibility in transforming the video summarization into a text summary task and leverage Large Language Models (LLMs) to boost video summarization. This paper proposes a novel self-supervised framework for video summarization guided by LLMs. Our method begins by generating captions for video frames, which are then synthesized into text summaries by LLMs. Subsequently, we measure semantic distance between the frame captions and the text summary. It's worth noting that we propose a novel loss function to optimize our model according to the diversity of the video. Finally, the summarized video can be generated by selecting the frames whose captions are similar with the text summary. Our model achieves competitive results against other state-of-the-art methods and paves a novel pathway in video summarization.

In the realm of autonomous driving, robust perception under out-of-distribution conditions is paramount for the safe deployment of vehicles. Challenges such as adverse weather, sensor malfunctions, and environmental unpredictability can severely impact the performance of autonomous systems. The 2024 RoboDrive Challenge was crafted to propel the development of driving perception technologies that can withstand and adapt to these real-world variabilities. Focusing on four pivotal tasks -- BEV detection, map segmentation, semantic occupancy prediction, and multi-view depth estimation -- the competition laid down a gauntlet to innovate and enhance system resilience against typical and atypical disturbances. This year's challenge consisted of five distinct tracks and attracted 140 registered teams from 93 institutes across 11 countries, resulting in nearly one thousand submissions evaluated through our servers. The competition culminated in 15 top-performing solutions, which introduced a range of innovative approaches including advanced data augmentation, multi-sensor fusion, self-supervised learning for error correction, and new algorithmic strategies to enhance sensor robustness. These contributions significantly advanced the state of the art, particularly in handling sensor inconsistencies and environmental variability. Participants, through collaborative efforts, pushed the boundaries of current technologies, showcasing their potential in real-world scenarios. Extensive evaluations and analyses provided insights into the effectiveness of these solutions, highlighting key trends and successful strategies for improving the resilience of driving perception systems. This challenge has set a new benchmark in the field, providing a rich repository of techniques expected to guide future research in this field.

We propose a simple strategy for masking image patches during visual-language contrastive learning that improves the quality of the learned representations and the training speed. During each iteration of training, we randomly mask clusters of visually similar image patches, as measured by their raw pixel intensities. This provides an extra learning signal, beyond the contrastive training itself, since it forces a model to predict words for masked visual structures solely from context. It also speeds up training by reducing the amount of data used in each image. We evaluate the effectiveness of our model by pre-training on a number of benchmarks, finding that it outperforms other masking strategies, such as FLIP, on the quality of the learned representation.

Current datasets for long-form video understanding often fall short of providing genuine long-form comprehension challenges, as many tasks derived from these datasets can be successfully tackled by analyzing just one or a few random frames from a video. To address this issue, we present a novel dataset and benchmark, CinePile, specifically designed for authentic long-form video understanding. This paper details our innovative approach for creating a question-answer dataset, utilizing advanced LLMs with human-in-the-loop and building upon human-generated raw data. Our comprehensive dataset comprises 305,000 multiple-choice questions (MCQs), covering various visual and multimodal aspects, including temporal comprehension, understanding human-object interactions, and reasoning about events or actions within a scene. Additionally, we evaluate recent video-centric LLMs, both open-source and proprietary, on the test split of our dataset. The findings reveal that even state-of-the-art video-centric LLMs significantly lag behind human performance in these tasks, highlighting the complexity and challenge inherent in video understanding. The dataset is available at https://hf.co/datasets/tomg-group-umd/cinepile

Large multimodal models (LMMs) have proven flexible and generalisable across many tasks and fields. Although they have strong potential to aid scientific research, their capabilities in this domain are not well characterised. A key aspect of scientific research is the ability to understand and interpret figures, which serve as a rich, compressed source of complex information. In this work, we present SciFIBench, a scientific figure interpretation benchmark. Our main benchmark consists of a 1000-question gold set of multiple-choice questions split between two tasks across 12 categories. The questions are curated from CS arXiv paper figures and captions, using adversarial filtering to find hard negatives and human verification for quality control. We evaluate 26 LMMs on SciFIBench, finding it to be a challenging benchmark. Finally, we investigate the alignment and reasoning faithfulness of the LMMs on augmented question sets from our benchmark. We release SciFIBench to encourage progress in this domain.

Datasets labelled by human annotators are widely used in the training and testing of machine learning models. In recent years, researchers are increasingly paying attention to label quality. However, it is not always possible to objectively determine whether an assigned label is correct or not. The present work investigates this ambiguity in the annotation of autonomous driving datasets as an important dimension of data quality. Our experiments show that excluding highly ambiguous data from the training improves model performance of a state-of-the-art pedestrian detector in terms of LAMR, precision and F1 score, thereby saving training time and annotation costs. Furthermore, we demonstrate that, in order to safely remove ambiguous instances and ensure the retained representativeness of the training data, an understanding of the properties of the dataset and class under investigation is crucial.

The Prostate Imaging Reporting and Data System (PI-RADS) is pivotal in the diagnosis of clinically significant prostate cancer through MRI imaging. Current deep learning-based PI-RADS scoring methods often lack the incorporation of essential PI-RADS clinical guidelines~(PICG) utilized by radiologists, potentially compromising scoring accuracy. This paper introduces a novel approach that adapts a multi-modal large language model (MLLM) to incorporate PICG into PI-RADS scoring without additional annotations and network parameters. We present a two-stage fine-tuning process aimed at adapting MLLMs originally trained on natural images to the MRI data domain while effectively integrating the PICG. In the first stage, we develop a domain adapter layer specifically tailored for processing 3D MRI image inputs and design the MLLM instructions to differentiate MRI modalities effectively. In the second stage, we translate PICG into guiding instructions for the model to generate PICG-guided image features. Through feature distillation, we align scoring network features with the PICG-guided image feature, enabling the scoring network to effectively incorporate the PICG information. We develop our model on a public dataset and evaluate it in a real-world challenging in-house dataset. Experimental results demonstrate that our approach improves the performance of current scoring networks.

Deep learning has enabled breakthroughs in automated diagnosis from medical imaging, with many successful applications in ophthalmology. However, standard medical image classification approaches only assess disease presence at the time of acquisition, neglecting the common clinical setting of longitudinal imaging. For slow, progressive eye diseases like age-related macular degeneration (AMD) and primary open-angle glaucoma (POAG), patients undergo repeated imaging over time to track disease progression and forecasting the future risk of developing disease is critical to properly plan treatment. Our proposed Longitudinal Transformer for Survival Analysis (LTSA) enables dynamic disease prognosis from longitudinal medical imaging, modeling the time to disease from sequences of fundus photography images captured over long, irregular time periods. Using longitudinal imaging data from the Age-Related Eye Disease Study (AREDS) and Ocular Hypertension Treatment Study (OHTS), LTSA significantly outperformed a single-image baseline in 19/20 head-to-head comparisons on late AMD prognosis and 18/20 comparisons on POAG prognosis. A temporal attention analysis also suggested that, while the most recent image is typically the most influential, prior imaging still provides additional prognostic value.

Indian folk paintings have a rich mosaic of symbols, colors, textures, and stories making them an invaluable repository of cultural legacy. The paper presents a novel approach to classifying these paintings into distinct art forms and tagging them with their unique salient features. A custom dataset named FolkTalent, comprising 2279 digital images of paintings across 12 different forms, has been prepared using websites that are direct outlets of Indian folk paintings. Tags covering a wide range of attributes like color, theme, artistic style, and patterns are generated using GPT4, and verified by an expert for each painting. Classification is performed employing the RandomForest ensemble technique on fine-tuned Convolutional Neural Network (CNN) models to classify Indian folk paintings, achieving an accuracy of 91.83%. Tagging is accomplished via the prominent fine-tuned CNN-based backbones with a custom classifier attached to its top to perform multi-label image classification. The generated tags offer a deeper insight into the painting, enabling an enhanced search experience based on theme and visual attributes. The proposed hybrid model sets a new benchmark in folk painting classification and tagging, significantly contributing to cataloging India's folk-art heritage.

The superior performance of modern visual backbones usually comes with a costly training procedure. We contribute to this issue by generalizing the idea of curriculum learning beyond its original formulation, i.e., training models using easier-to-harder data. Specifically, we reformulate the training curriculum as a soft-selection function, which uncovers progressively more difficult patterns within each example during training, instead of performing easier-to-harder sample selection. Our work is inspired by an intriguing observation on the learning dynamics of visual backbones: during the earlier stages of training, the model predominantly learns to recognize some 'easier-to-learn' discriminative patterns in the data. These patterns, when observed through frequency and spatial domains, incorporate lower-frequency components, and the natural image contents without distortion or data augmentation. Motivated by these findings, we propose a curriculum where the model always leverages all the training data at every learning stage, yet the exposure to the 'easier-to-learn' patterns of each example is initiated first, with harder patterns gradually introduced as training progresses. To implement this idea in a computationally efficient way, we introduce a cropping operation in the Fourier spectrum of the inputs, enabling the model to learn from only the lower-frequency components. Then we show that exposing the contents of natural images can be readily achieved by modulating the intensity of data augmentation. Finally, we integrate these aspects and design curriculum schedules with tailored search algorithms. The resulting method, EfficientTrain++, is simple, general, yet surprisingly effective. It reduces the training time of a wide variety of popular models by 1.5-3.0x on ImageNet-1K/22K without sacrificing accuracy. It also demonstrates efficacy in self-supervised learning (e.g., MAE).

Out-of-distribution (OOD) detection is critical when deploying machine learning models in the real world. Outlier exposure methods, which incorporate auxiliary outlier data in the training process, can drastically improve OOD detection performance compared to approaches without advanced training strategies. We introduce Hopfield Boosting, a boosting approach, which leverages modern Hopfield energy (MHE) to sharpen the decision boundary between the in-distribution and OOD data. Hopfield Boosting encourages the model to concentrate on hard-to-distinguish auxiliary outlier examples that lie close to the decision boundary between in-distribution and auxiliary outlier data. Our method achieves a new state-of-the-art in OOD detection with outlier exposure, improving the FPR95 metric from 2.28 to 0.92 on CIFAR-10 and from 11.76 to 7.94 on CIFAR-100.

Few-shot segmentation (FSS) aims to train a model which can segment the object from novel classes with a few labeled samples. The insufficient generalization ability of models leads to unsatisfactory performance when the models lack enough labeled data from the novel classes. Considering that there are abundant unlabeled data available, it is promising to improve the generalization ability by exploiting these various data. For leveraging unlabeled data, we propose a novel method, named Image to Pseudo-Episode (IPE), to generate pseudo-episodes from unlabeled data. Specifically, our method contains two modules, i.e., the pseudo-label generation module and the episode generation module. The former module generates pseudo-labels from unlabeled images by the spectral clustering algorithm, and the latter module generates pseudo-episodes from pseudo-labeled images by data augmentation methods. Extensive experiments on PASCAL-$5^i$ and COCO-$20^i$ demonstrate that our method achieves the state-of-the-art performance for FSS.

We present Hunyuan-DiT, a text-to-image diffusion transformer with fine-grained understanding of both English and Chinese. To construct Hunyuan-DiT, we carefully design the transformer structure, text encoder, and positional encoding. We also build from scratch a whole data pipeline to update and evaluate data for iterative model optimization. For fine-grained language understanding, we train a Multimodal Large Language Model to refine the captions of the images. Finally, Hunyuan-DiT can perform multi-turn multimodal dialogue with users, generating and refining images according to the context. Through our holistic human evaluation protocol with more than 50 professional human evaluators, Hunyuan-DiT sets a new state-of-the-art in Chinese-to-image generation compared with other open-source models. Code and pretrained models are publicly available at github.com/Tencent/HunyuanDiT

In this paper, we present a simple but effective method to enhance blind video quality assessment (BVQA) models for social media videos. Motivated by previous researches that leverage pre-trained features extracted from various computer vision models as the feature representation for BVQA, we further explore rich quality-aware features from pre-trained blind image quality assessment (BIQA) and BVQA models as auxiliary features to help the BVQA model to handle complex distortions and diverse content of social media videos. Specifically, we use SimpleVQA, a BVQA model that consists of a trainable Swin Transformer-B and a fixed SlowFast, as our base model. The Swin Transformer-B and SlowFast components are responsible for extracting spatial and motion features, respectively. Then, we extract three kinds of features from Q-Align, LIQE, and FAST-VQA to capture frame-level quality-aware features, frame-level quality-aware along with scene-specific features, and spatiotemporal quality-aware features, respectively. Through concatenating these features, we employ a multi-layer perceptron (MLP) network to regress them into quality scores. Experimental results demonstrate that the proposed model achieves the best performance on three public social media VQA datasets. Moreover, the proposed model won first place in the CVPR NTIRE 2024 Short-form UGC Video Quality Assessment Challenge. The code is available at \url{https://github.com/sunwei925/RQ-VQA.git}.

We present a simple algorithm for differentiable rendering of surfaces represented by Signed Distance Fields (SDF), which makes it easy to integrate rendering into gradient-based optimization pipelines. To tackle visibility-related derivatives that make rendering non-differentiable, existing physically based differentiable rendering methods often rely on elaborate guiding data structures or reparameterization with a global impact on variance. In this article, we investigate an alternative that embraces nonzero bias in exchange for low variance and architectural simplicity. Our method expands the lower-dimensional boundary integral into a thin band that is easy to sample when the underlying surface is represented by an SDF. We demonstrate the performance and robustness of our formulation in end-to-end inverse rendering tasks, where it obtains results that are competitive with or superior to existing work.

Addressing multi-label action recognition in videos represents a significant challenge for robotic applications in dynamic environments, especially when the robot is required to cooperate with humans in tasks that involve objects. Existing methods still struggle to recognize unseen actions or require extensive training data. To overcome these problems, we propose Dual-VCLIP, a unified approach for zero-shot multi-label action recognition. Dual-VCLIP enhances VCLIP, a zero-shot action recognition method, with the DualCoOp method for multi-label image classification. The strength of our method is that at training time it only learns two prompts, and it is therefore much simpler than other methods. We validate our method on the Charades dataset that includes a majority of object-based actions, demonstrating that -- despite its simplicity -- our method performs favorably with respect to existing methods on the complete dataset, and promising performance when tested on unseen actions. Our contribution emphasizes the impact of verb-object class-splits during robots' training for new cooperative tasks, highlighting the influence on the performance and giving insights into mitigating biases.

Numerous studies have revealed that deep learning-based medical image classification models may exhibit bias towards specific demographic attributes, such as race, gender, and age. Existing bias mitigation methods often achieve high level of fairness at the cost of significant accuracy degradation. In response to this challenge, we propose an innovative and adaptable Soft Nearest Neighbor Loss-based channel pruning framework, which achieves fairness through channel pruning. Traditionally, channel pruning is utilized to accelerate neural network inference. However, our work demonstrates that pruning can also be a potent tool for achieving fairness. Our key insight is that different channels in a layer contribute differently to the accuracy of different groups. By selectively pruning critical channels that lead to the accuracy difference between the privileged and unprivileged groups, we can effectively improve fairness without sacrificing accuracy significantly. Experiments conducted on two skin lesion diagnosis datasets across multiple sensitive attributes validate the effectiveness of our method in achieving state-of-the-art trade-off between accuracy and fairness. Our code is available at https://github.com/Kqp1227/Sensitive-Channel-Pruning.

Depth estimation plays a crucial role in various tasks within endoscopic surgery, including navigation, surface reconstruction, and augmented reality visualization. Despite the significant achievements of foundation models in vision tasks, including depth estimation, their direct application to the medical domain often results in suboptimal performance. This highlights the need for efficient adaptation methods to adapt these models to endoscopic depth estimation. We propose Endoscopic Depth Any Camera (EndoDAC) which is an efficient self-supervised depth estimation framework that adapts foundation models to endoscopic scenes. Specifically, we develop the Dynamic Vector-Based Low-Rank Adaptation (DV-LoRA) and employ Convolutional Neck blocks to tailor the foundational model to the surgical domain, utilizing remarkably few trainable parameters. Given that camera information is not always accessible, we also introduce a self-supervised adaptation strategy that estimates camera intrinsics using the pose encoder. Our framework is capable of being trained solely on monocular surgical videos from any camera, ensuring minimal training costs. Experiments demonstrate that our approach obtains superior performance even with fewer training epochs and unaware of the ground truth camera intrinsics. Code is available at https://github.com/BeileiCui/EndoDAC.

Large-scale Vision-Language Models (VLMs) have demonstrated exceptional performance in natural vision tasks, motivating researchers across domains to explore domain-specific VLMs. However, the construction of powerful domain-specific VLMs demands vast amounts of annotated data, substantial electrical energy, and computing resources, primarily accessible to industry, yet hindering VLM research in academia. To address this challenge and foster sustainable and equitable VLM research, we present the Generalized Domain Prompt Learning (GDPL) framework. GDPL facilitates the transfer of VLMs' robust recognition capabilities from natural vision to specialized domains, without the need for extensive data or resources. By leveraging small-scale domain-specific foundation models and minimal prompt samples, GDPL empowers the language branch with domain knowledge through quaternion networks, uncovering cross-modal relationships between domain-specific vision features and natural vision-based contextual embeddings. Simultaneously, GDPL guides the vision branch into specific domains through hierarchical propagation of generated vision prompt features, grounded in well-matched vision-language relations. Furthermore, to fully harness the domain adaptation potential of VLMs, we introduce a novel low-rank adaptation approach. Extensive experiments across diverse domains like remote sensing, medical imaging, geology, Synthetic Aperture Radar, and fluid dynamics, validate the efficacy of GDPL, demonstrating its ability to achieve state-of-the-art domain recognition performance in a prompt learning paradigm. Our framework paves the way for sustainable and inclusive VLM research, transcending the barriers between academia and industry.

The increasing complexity of Artificial Intelligence models poses challenges to interpretability, particularly in the healthcare sector. This study investigates the impact of deep learning model complexity and Explainable AI (XAI) efficacy, utilizing four ResNet architectures (ResNet-18, 34, 50, 101). Through methodical experimentation on 4,369 lung X-ray images of COVID-19-infected and healthy patients, the research evaluates models' classification performance and the relevance of corresponding XAI explanations with respect to the ground-truth disease masks. Results indicate that the increase in model complexity is associated with a decrease in classification accuracy and AUC-ROC scores (ResNet-18: 98.4%, 0.997; ResNet-101: 95.9%, 0.988). Notably, in eleven out of twelve statistical tests performed, no statistically significant differences occurred between XAI quantitative metrics - Relevance Rank Accuracy and the proposed Positive Attribution Ratio - across trained models. These results suggest that increased model complexity does not consistently lead to higher performance or relevance of explanations for models' decision-making processes.

Self-supervised learning (SSL) is an approach to extract useful feature representations from unlabeled data, and enable fine-tuning on downstream tasks with limited labeled examples. Self-pretraining is a SSL approach that uses the curated task dataset for both pretraining the networks and fine-tuning them. Availability of large, diverse, and uncurated public medical image sets provides the opportunity to apply SSL in the "wild" and potentially extract features robust to imaging variations. However, the benefit of wild- vs self-pretraining has not been studied for medical image analysis. In this paper, we compare robustness of wild versus self-pretrained transformer (vision transformer [ViT] and hierarchical shifted window [Swin]) models to computed tomography (CT) imaging differences for non-small cell lung cancer (NSCLC) segmentation. Wild-pretrained Swin models outperformed self-pretrained Swin for the various imaging acquisitions. ViT resulted in similar accuracy for both wild- and self-pretrained models. Masked image prediction pretext task that forces networks to learn the local structure resulted in higher accuracy compared to contrastive task that models global image information. Wild-pretrained models resulted in higher feature reuse at the lower level layers and feature differentiation close to output layer after fine-tuning. Hence, we conclude: Wild-pretrained networks were more robust to analyzed CT imaging differences for lung tumor segmentation than self-pretrained methods. Swin architecture benefited from such pretraining more than ViT.

With the increasing use of neural networks in critical systems, runtime monitoring becomes essential to reject unsafe predictions during inference. Various techniques have emerged to establish rejection scores that maximize the separability between the distributions of safe and unsafe predictions. The efficacy of these approaches is mostly evaluated using threshold-agnostic metrics, such as the area under the receiver operating characteristic curve. However, in real-world applications, an effective monitor also requires identifying a good threshold to transform these scores into meaningful binary decisions. Despite the pivotal importance of threshold optimization, this problem has received little attention. A few studies touch upon this question, but they typically assume that the runtime data distribution mirrors the training distribution, which is a strong assumption as monitors are supposed to safeguard a system against potentially unforeseen threats. In this work, we present rigorous experiments on various image datasets to investigate: 1. The effectiveness of monitors in handling unforeseen threats, which are not available during threshold adjustments. 2. Whether integrating generic threats into the threshold optimization scheme can enhance the robustness of monitors.

Neural Radiance Field(NeRF) is an novel implicit method to achieve the 3D reconstruction and representation with a high resolution. After the first research of NeRF is proposed, NeRF has gained a robust developing power and is booming in the 3D modeling, representation and reconstruction areas. However the first and most of the followed research projects based on NeRF is static, which are weak in the practical applications. Therefore, more researcher are interested and focused on the study of dynamic NeRF that is more feasible and useful in practical applications or situations. Compared with the static NeRF, implementing the Dynamic NeRF is more difficult and complex. But Dynamic is more potential in the future even is the basic of Editable NeRF. In this review, we made a detailed and abundant statement for the development and important implementation principles of Dynamci NeRF. The analysis of main principle and development of Dynamic NeRF is from 2021 to 2023, including the most of the Dynamic NeRF projects. What is more, with colorful and novel special designed figures and table, We also made a detailed comparison and analysis of different features of various of Dynamic. Besides, we analyzed and discussed the key methods to implement a Dynamic NeRF. The volume of the reference papers is large. The statements and comparisons are multidimensional. With a reading of this review, the whole development history and most of the main design method or principles of Dynamic NeRF can be easy understood and gained.

The nature of diversity in real-world environments necessitates neural network models to expand from closed category settings to accommodate novel emerging categories. In this paper, we study the open-vocabulary object detection (OVD), which facilitates the detection of novel object classes under the supervision of only base annotations and open-vocabulary knowledge. However, we find that the inadequacy of neighboring relationships between regions during the alignment process inevitably constrains the performance on recent distillation-based OVD strategies. To this end, we propose Neighboring Region Attention Alignment (NRAA), which performs alignment within the attention mechanism of a set of neighboring regions to boost the open-vocabulary inference. Specifically, for a given proposal region, we randomly explore the neighboring boxes and conduct our proposed neighboring region attention (NRA) mechanism to extract relationship information. Then, this interaction information is seamlessly provided into the distillation procedure to assist the alignment between the detector and the pre-trained vision-language models (VLMs). Extensive experiments validate that our proposed model exhibits superior performance on open-vocabulary benchmarks.

In many applications, the demand arises for algorithms capable of aligning partially overlapping point sets while remaining invariant to the corresponding transformations. This research presents a method designed to meet such requirements through minimization of the objective function of the robust point matching (RPM) algorithm. First, we show that the RPM objective is a cubic polynomial. Then, through variable substitution, we transform the RPM objective to a quadratic function. Leveraging the convex envelope of bilinear monomials, we proceed to relax the resulting objective function, thus obtaining a lower bound problem that can be conveniently decomposed into distinct linear assignment and low-dimensional convex quadratic program components, both amenable to efficient optimization. Furthermore, a branch-and-bound (BnB) algorithm is devised, which solely branches over the transformation parameters, thereby boosting convergence rate. Empirical evaluations demonstrate better robustness of the proposed methodology against non-rigid deformation, positional noise, and outliers, particularly in scenarios where outliers remain distinct from inliers, when compared with prevailing state-of-the-art approaches.

Tissue tracking in echocardiography is challenging due to the complex cardiac motion and the inherent nature of ultrasound acquisitions. Although optical flow methods are considered state-of-the-art (SOTA), they struggle with long-range tracking, noise occlusions, and drift throughout the cardiac cycle. Recently, novel learning-based point tracking techniques have been introduced to tackle some of these issues. In this paper, we build upon these techniques and introduce EchoTracker, a two-fold coarse-to-fine model that facilitates the tracking of queried points on a tissue surface across ultrasound image sequences. The architecture contains a preliminary coarse initialization of the trajectories, followed by reinforcement iterations based on fine-grained appearance changes. It is efficient, light, and can run on mid-range GPUs. Experiments demonstrate that the model outperforms SOTA methods, with an average position accuracy of 67% and a median trajectory error of 2.86 pixels. Furthermore, we show a relative improvement of 25% when using our model to calculate the global longitudinal strain (GLS) in a clinical test-retest dataset compared to other methods. This implies that learning-based point tracking can potentially improve performance and yield a higher diagnostic and prognostic value for clinical measurements than current techniques. Our source code is available at: https://github.com/riponazad/echotracker/.

Domain generalization aims to develop models that are robust to distribution shifts. Existing methods focus on learning invariance across domains to enhance model robustness, and data augmentation has been widely used to learn invariant predictors, with most methods performing augmentation in the input space. However, augmentation in the input space has limited diversity whereas in the feature space is more versatile and has shown promising results. Nonetheless, feature semantics is seldom considered and existing feature augmentation methods suffer from a limited variety of augmented features. We decompose features into class-generic, class-specific, domain-generic, and domain-specific components. We propose a cross-domain feature augmentation method named XDomainMix that enables us to increase sample diversity while emphasizing the learning of invariant representations to achieve domain generalization. Experiments on widely used benchmark datasets demonstrate that our proposed method is able to achieve state-of-the-art performance. Quantitative analysis indicates that our feature augmentation approach facilitates the learning of effective models that are invariant across different domains.

Image stitching aims to construct a wide field of view with high spatial resolution, which cannot be achieved in a single exposure. Typically, conventional image stitching techniques, other than deep learning, require complex computation and thus computational pricy, especially for stitching large raw images. In this study, inspired by the multiscale feature of fluid turbulence, we developed a fast feature point detection algorithm named local-peak scale-invariant feature transform (LP-SIFT), based on the multiscale local peaks and scale-invariant feature transform method. By combining LP-SIFT and RANSAC in image stitching, the stitching speed can be improved by orders, compared with the original SIFT method. Nine large images (over 2600*1600 pixels), arranged randomly without prior knowledge, can be stitched within 158.94 s. The algorithm is highly practical for applications requiring a wide field of view in diverse application scenes, e.g., terrain mapping, biological analysis, and even criminal investigation.

Although pre-training on a large amount of data is beneficial for robot learning, current paradigms only perform large-scale pretraining for visual representations, whereas representations for other modalities are trained from scratch. In contrast to the abundance of visual data, it is unclear what relevant internet-scale data may be used for pretraining other modalities such as tactile sensing. Such pretraining becomes increasingly crucial in the low-data regimes common in robotics applications. In this paper, we address this gap by using contact microphones as an alternative tactile sensor. Our key insight is that contact microphones capture inherently audio-based information, allowing us to leverage large-scale audio-visual pretraining to obtain representations that boost the performance of robotic manipulation. To the best of our knowledge, our method is the first approach leveraging large-scale multisensory pre-training for robotic manipulation. For supplementary information including videos of real robot experiments, please see https://sites.google.com/view/hearing-touch.

This paper addresses the problem of pathological lung segmentation, a significant challenge in medical image analysis, particularly pronounced in cases of peripheral opacities (severe fibrosis and consolidation) because of the textural similarity between lung tissue and surrounding areas. To overcome these challenges, this paper emphasizes the use of CycleGAN for unpaired image-to-image translation, in order to provide an augmentation method able to generate fake pathological images matching an existing ground truth. Although previous studies have employed CycleGAN, they often neglect the challenge of shape deformation, which is crucial for accurate medical image segmentation. Our work introduces an innovative strategy that incorporates additional loss functions. Specifically, it proposes an L1 loss based on the lung surrounding which shape is constrained to remain unchanged at the transition from the healthy to pathological domains. The lung surrounding is derived based on ground truth lung masks available in the healthy domain. Furthermore, preprocessing steps, such as cropping based on ribs/vertebra locations, are applied to refine the input for the CycleGAN, ensuring that the network focus on the lung region. This is essential to avoid extraneous biases, such as the zoom effect bias, which can divert attention from the main task. The method is applied to enhance in semi-supervised manner the lung segmentation process by employing a U-Net model trained with on-the-fly data augmentation incorporating synthetic pathological tissues generated by the CycleGAN model. Preliminary results from this research demonstrate significant qualitative and quantitative improvements, setting a new benchmark in the field of pathological lung segmentation. Our code is available at https://github.com/noureddinekhiati/Semi-supervised-lung-segmentation

Portrait images typically consist of a salient person against diverse backgrounds. With the development of mobile devices and image processing techniques, users can conveniently capture portrait images anytime and anywhere. However, the quality of these portraits may suffer from the degradation caused by unfavorable environmental conditions, subpar photography techniques, and inferior capturing devices. In this paper, we introduce a dual-branch network for portrait image quality assessment (PIQA), which can effectively address how the salient person and the background of a portrait image influence its visual quality. Specifically, we utilize two backbone networks (\textit{i.e.,} Swin Transformer-B) to extract the quality-aware features from the entire portrait image and the facial image cropped from it. To enhance the quality-aware feature representation of the backbones, we pre-train them on the large-scale video quality assessment dataset LSVQ and the large-scale facial image quality assessment dataset GFIQA. Additionally, we leverage LIQE, an image scene classification and quality assessment model, to capture the quality-aware and scene-specific features as the auxiliary features. Finally, we concatenate these features and regress them into quality scores via a multi-perception layer (MLP). We employ the fidelity loss to train the model via a learning-to-rank manner to mitigate inconsistencies in quality scores in the portrait image quality assessment dataset PIQ. Experimental results demonstrate that the proposed model achieves superior performance in the PIQ dataset, validating its effectiveness. The code is available at \url{https://github.com/sunwei925/DN-PIQA.git}.

Class-incremental learning (CIL) has emerged as a means to learn new classes incrementally without catastrophic forgetting of previous classes. Recently, CIL has undergone a paradigm shift towards dynamic architectures due to their superior performance. However, these models are still limited by the following aspects: (i) Data augmentation (DA), which are tightly coupled with CIL, remains under-explored in dynamic architecture scenarios. (ii) Feature representation. The discriminativeness of dynamic feature are sub-optimal and possess potential for refinement. (iii) Classifier. The misalignment between dynamic feature and classifier constrains the capabilities of the model. To tackle the aforementioned drawbacks, we propose the Dynamic Feature Learning and Matching (DFLM) model in this paper from above three perspectives. Specifically, we firstly introduce class weight information and non-stationary functions to extend the mix DA method for dynamically adjusting the focus on memory during training. Then, von Mises-Fisher (vMF) classifier is employed to effectively model the dynamic feature distribution and implicitly learn their discriminative properties. Finally, the matching loss is proposed to facilitate the alignment between the learned dynamic features and the classifier by minimizing the distribution distance. Extensive experiments on CIL benchmarks validate that our proposed model achieves significant performance improvements over existing methods.

Deep learning methods, especially Convolutional Neural Networks (CNN) and Vision Transformer (ViT), are frequently employed to perform semantic segmentation of high-resolution remotely sensed images. However, CNNs are constrained by their restricted receptive fields, while ViTs face challenges due to their quadratic complexity. Recently, the Mamba model, featuring linear complexity and a global receptive field, has gained extensive attention for vision tasks. In such tasks, images need to be serialized to form sequences compatible with the Mamba model. Numerous research efforts have explored scanning strategies to serialize images, aiming to enhance the Mamba model's understanding of images. However, the effectiveness of these scanning strategies remains uncertain. In this research, we conduct a comprehensive experimental investigation on the impact of mainstream scanning directions and their combinations on semantic segmentation of remotely sensed images. Through extensive experiments on the LoveDA, ISPRS Potsdam, and ISPRS Vaihingen datasets, we demonstrate that no single scanning strategy outperforms others, regardless of their complexity or the number of scanning directions involved. A simple, single scanning direction is deemed sufficient for semantic segmentation of high-resolution remotely sensed images. Relevant directions for future research are also recommended.

In recent years, deep learning has greatly streamlined the process of generating realistic fake face images. Aware of the dangers, researchers have developed various tools to spot these counterfeits. Yet none asked the fundamental question: What digital manipulations make a real photographic face image fake, while others do not? In this paper, we put face forgery in a semantic context and define that computational methods that alter semantic face attributes to exceed human discrimination thresholds are sources of face forgery. Guided by our new definition, we construct a large face forgery image dataset, where each image is associated with a set of labels organized in a hierarchical graph. Our dataset enables two new testing protocols to probe the generalization of face forgery detectors. Moreover, we propose a semantics-oriented face forgery detection method that captures label relations and prioritizes the primary task (\ie, real or fake face detection). We show that the proposed dataset successfully exposes the weaknesses of current detectors as the test set and consistently improves their generalizability as the training set. Additionally, we demonstrate the superiority of our semantics-oriented method over traditional binary and multi-class classification-based detectors.