论文检索

Huan Li, Hailong Shi, Xingyu Gao

Event-driven image deblurring is an innovative approach involving the input of events obtained from the event camera alongside blurred frames to facilitate the deblurring process. Unlike conventional cameras, event cameras in event-driven imaging exhibit low-latency characteristics and are immune to motion blur, resulting in significant advancements in image deblurring. In this paper, we propose a pioneering event-based coarse-to-fine image deblurring network named CFFNet. In contrast to existing deblurring methods, our approach incorporates event data, generating multiple coarse frames from a single frame before further refining them into a sharp image. We introduce an Event Image Fusion Block (EIFB) for the coarse fusion of events and images, producing coarse frames at different time points. Additionally, we propose a Bidirectional Frame Fusion Block (BFFB) for the fine fusion of coarse frames. CFFNet effectively harnesses the spatiotemporal information of event data through a comprehensive fusion process from coarse to fine. Experimental results on the GoPro and REBlur datasets demonstrate that our method achieves state-of-the-art performance for image deblurring task.

Bohan Li, Yasheng Sun, Zhujin Liang, Dalong Du, Zhuanghui Zhang, Xiaofeng Wang, Yunnan Wang, Xin Jin, Wenjun Zeng

3D semantic scene completion (SSC) is an ill-posed perception task that requires inferring a dense 3D scene from limited observations. Previous camera-based methods struggle to predict accurate semantic scenes due to inherent geometric ambiguity and incomplete observations. In this paper, we resort to stereo matching technique and bird’s-eye-view (BEV) representation learning to address such issues in SSC. Complementary to each other, stereo matching mitigates geometric ambiguity with epipolar constraint while BEV representation enhances the hallucination ability for invisible regions with global semantic context. However, due to the inherent representation gap between stereo geometry and BEV features, it is non-trivial to bridge them for dense prediction task of SSC. Therefore, we further develop a unified occupancy-based framework dubbed BRGScene, which effectively bridges these two representations with dense 3D volumes for reliable semantic scene completion. Specifically, we design a novel Mutual Interactive Ensemble (MIE) block for pixel-level reliable aggregation of stereo geometry and BEV features. Within the MIE block, a Bi-directional Reliable Interaction (BRI) module, enhanced with confidence re-weighting, is employed to encourage fine-grained interaction through mutual guidance. Besides, a Dual Volume Ensemble (DVE) module is introduced to facilitate complementary aggregation through channel-wise recalibration and multi-group voting. Our method outperforms all published camera-based methods on SemanticKITTI for semantic scene completion. Our code is available on https://github.com/Arlo0o/StereoScene.

Eun-Gi Lee, Moon Seok Lee, Jae Hyun Yoon, Seok Bong Yoo

Adversarial attacks pose a severe threat to the accuracy of person re-identification (re-ID) systems, a critical security technology. Adversarial purification methods are promising approaches for defending against comprehensive attacks, including unseen ones. However, re-ID testing identities (IDs) are unseen, requiring more sophisticated purification than other classification tasks for adversarial defense. We propose IntensPure, an adversarial purification method in person re-ID that quantifies attack intensity via ID stability and attribute inconsistency to customize purification strength. Based on the estimated attack intensity, IntensPure employs secondary domain adaptive diffusion focused on purifying the low- and mid-frequency coefficients vulnerable to re-ID attacks. This method significantly reduces computational costs compared to the conventional diffusion method. For elaborate purification, IntensPure performs a directional diffusion process and refinements, leveraging the directional characteristics of secondary images. The experimental results on diverse attacks demonstrate that IntensPure outperforms the existing methods in terms of rank-1 accuracy.

Minsoo Khang, Teakgyu Hong

Table Structure Recognition (TSR) is a task aimed at converting table images into a machine-readable format (e.g. HTML), to facilitate other applications such as information retrieval. Recent works tackle this problem by identifying the HTML tags and text regions, where the latter is used for text extraction from the table document. These works however, suffer from misalignment issues when mapping text into the identified text regions. In this paper, we introduce a new TSR framework, called TFLOP (TSR Framework with LayOut Pointer mechanism), which reformulates the conventional text region prediction and matching into a direct text region pointing problem. Specifically, TFLOP utilizes text region information to identify both the table's structure tags and its aligned text regions, simultaneously. Without the need for region prediction and alignment, TFLOP circumvents the additional text region matching stage, which requires finely-calibrated post-processing. TFLOP also employs span-aware contrastive supervision to enhance the pointing mechanism in tables with complex structure. As a result, TFLOP achieves the state-of-the-art performance across multiple benchmarks such as PubTabNet, FinTabNet, and SynthTabNet. In our extensive experiments, TFLOP not only exhibits competitive performance but also shows promising results on industrial document TSR scenarios such as documents with watermarks or in non-English domain. Source code of our work is publicly available at: https://github.com/UpstageAI/TFLOP.

Yimin Jiang, Huibing Wang, Jinjia Peng, Xianping Fu, Yang Wang

Person search aims to localize specific a target person from a gallery set of images with various scenes. As the scene of moving pedestrian changes, the captured person image inevitably bring in lots of background noise and foreground noise on the person feature, which are completely unrelated to the person identity, leading to severe performance degeneration. To address this issue, we present a Scene-Adaptive Person Search (SEAS) model by introducing bilateral modulations to simultaneously eliminate scene noise and maintain a consistent person representation to adapt to various scenes. In SEAS, a Background Modulation Network (BMN) is designed to encode the feature extracted from the detected bounding box into a multi-granularity embedding, which reduces the input of background noise from multiple levels with norm-aware. Additionally, to mitigate the effect of foreground noise on the person feature, SEAS introduces a Foreground Modulation Network (FMN) to compute the clutter reduction offset for the person embedding based on the feature map of the scene image. By bilateral modulations on both background and foreground within an end-to-end manner, SEAS obtains consistent feature representations without scene noise. SEAS can achieve state-of-the-art (SOTA) performance on two benchmark datasets, CUHK-SYSU with 97.1% mAP and PRW with 60.5% mAP. The code is available at https://github.com/whbdmu/SEAS.

Wentao Jiang, Jing Zhang, Di Wang, Qiming Zhang, Zengmao Wang, Bo Du

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 × speedup, fewer parameters, and competitive performance compared to the baseline models, and achieves a better trade-off between efficiency and performance. The code is released at https://github.com/ViTAE-Transformer/LeMeViT.

Dongxiang Jiang, Yongchang Zhang, Shuai He, Anlong Ming

In recent years, short-form videos have gained popularity and the editing of these videos, particularly when motion is synchronized with music, is highly favored due to its beat-matching effect. However, detecting motion rhythm poses a significant challenge as it is influenced by multiple factors that make it difficult to define using explicit rules. While traditional methods attempt to define motion rhythm, they often yield unsatisfactory results. On the other hand, learning-based methods can extract motion rhythm without relying on explicit rules but require high-quality datasets. Unfortunately, existing datasets simply substitute music rhythm for motion rhythm which are not equivalent. To address these challenges, we present the motion rhythm dataset AIST-M2B, which is annotated with meticulously curated motion rhythm labels derived from the profound correlation between motion and music in professional dance. We propose a novel network architecture called M2BNet that is specifically trained on AIST-M2B to effectively extract intricate motion rhythms by incorporating both human body structure and temporal information. Additionally, we introduce a pioneering algorithm for enhancing motion rhythm synchronization with beats. Experimental results substan- tiate the superior performance of our method compared to other existing algorithms in the domain of motion rhythm analysis. Our code is available at https://github.com/mRobotit/M2Beats.

Yunbing Jia, Xiaoyu Kong, Fan Tang, Yixing Gao, Weiming Dong, Yi Yang

In this paper, we reveal the two sides of data augmentation: enhancements in closed-set recognition correlate with a significant decrease in open-set recognition. Through empirical investigation, we find that multi-sample-based augmentations would contribute to reducing feature discrimination, thereby diminishing the open-set criteria. Although knowledge distillation could impair the feature via imitation, the mixed feature with ambiguous semantics hinders the distillation. To this end, we propose an asymmetric distillation framework by feeding teacher model extra raw data to enlarge the benefit of teacher. Moreover, a joint mutual information loss and a selective relabel strategy are utilized to alleviate the influence of hard mixed samples. Our method successfully mitigates the decline in open-set and outperforms SOTAs by 2%~3% AUROC on the Tiny-ImageNet dataset and experiments on large-scale dataset ImageNet-21K demonstrate the generalization of our method.

Xiang Ji, Changqiao Xu, Lujie Zhong, Shujie Yang, Han Xiao, Gabriel-Miro Muntean

Omnidirectional images (ODIs) demand considerably higher resolution to ensure high quality across all viewports. Traditional convolutional neural networks (CNN)-based single-image super-resolution (SISR) networks, however, are not effective for spherical ODIs. This is due to the uneven pixel density distribution and varying texture complexity in different regions that arise when projecting from a sphere to a plane. Additionally, the computational and memory costs associated with large-sized ODIs present a challenge for real-world application. To address these issues, we propose an efficient distortion-adaptive super-resolution network (ODA-SRN). Specifically, ODA-SRN employs a series of specially designed Distortion Attention Block Groups (DABG) as its backbone. Our Distortion Attention Blocks (DABs) utilize multi-segment parameterized convolution to generate dynamic filters, which compensate for distortion and texture fading during feature extraction. Moreover, we introduce an upsampling scheme that accounts for the dependence of pixel position and distortion degree to achieve pixel-level distortion offset. A comprehensive set of results demonstrates that our ODA-SRN significantly improves the super-resolution performance for ODIs, both quantitatively and qualitatively, when compared to other state-of-the-art methods.

Deyi Ji, Wenwei Jin, Hongtao Lu, Feng Zhao

The ascension of Unmanned Aerial Vehicles (UAVs) in various fields necessitates effective UAV image segmentation, which faces challenges due to the dynamic perspectives of UAV-captured images. Traditional segmentation algorithms falter as they cannot accurately mimic the complexity of UAV perspectives, and the cost of obtaining multi-perspective labeled datasets is prohibitive. To address these issues, we introduce the PPTFormer, a novel Pseudo Multi-Perspective Transformer network that revolutionizes UAV image segmentation. Our approach circumvents the need for actual multi-perspective data by creating pseudo perspectives for enhanced multi-perspective learning. The PPTFormer network boasts Perspective Decomposition, novel Perspective Prototypes, and a specialized encoder and decoder that together achieve superior segmentation results through Pseudo Multi-Perspective Attention (PMP Attention) and fusion. Our experiments demonstrate that PPTFormer achieves state-of-the-art performance across five UAV segmentation datasets, confirming its capability to effectively simulate UAV flight perspectives and significantly advance segmentation precision. This work presents a pioneering leap in UAV scene understanding and sets a new benchmark for future developments in semantic segmentation.

Zi-Xian Huang, Chuan-Xian Ren

Open Set Domain Adaptation (OSDA) aims to transfer knowledge from a labeled source domain to an unlabeled target domain, where known classes exist across domains while unknown classes are present only in the target domain. Existing methods rely on the clustering structure to identify the unknown classes, which empirically induces a large identification error if the unknown classes are a mixture of multiple components. To break through this barrier, we formulate OSDA from the view of correlation and propose a correlation metric-based framework called Balanced Correlation Learning (BCL). BCL employs Hilbert-Schmidt Independence Criterion (HSIC) to characterize the separation between unknown and known classes, where HSIC is reformulated as the nodes’ relation on graph. By considering the label prior as variable, theoretical results are derived to analytically show a sufficient condition for desired learning direction for OSDA. Methodologically, the class-balanced HSIC is proposed to preserve domain-invariant and class-discriminative features. With the guarantee of correlation learning, the entropy-based principle can effectively identify the unknown classes via uncertainty. Empirically, extensive evaluations are conducted, where BCL achieves significant performance improvements.

Muqi Huang, Chaoyue Wang, Yong Luo, Lefei Zhang

Text-guided image inpainting has rapidly garnered prominence as a task in user-directed image synthesis, aiming to complete the occluded image regions following the textual prompt provided. However, current methods usually grapple with issues arising from the disparity between low-level pixel data and high-level semantic descriptions, which results in inpainted sections not harmonizing with the original image (either structurally or texturally). In this study, we introduce a Structure-Aware Inpainting Learning scheme and an Asymmetric Cross Domain Attention to address these cross-domain misalignment challenges. The proposed structure-aware learning scheme employs features of an intermediate modality as structure guidance to bridge the gap between text information and low-level pixels. Meanwhile, asymmetric cross-domain attention enhances the texture consistency between inpainted and unmasked regions. Our experiments show exceptional performance on leading datasets such as MS-COCO and Open Images, surpassing state-of-the-art text-guided image inpainting methods. Code is released at: https://github.com/MucciH/ECDM-inpainting

Chao Huang, Jie Wen, Chengliang Liu, Yabo Liu

Video anomaly detection (VAD) is the core problem of intelligent video surveillance. Previous methods commonly adopt the unsupervised paradigm of frame reconstruction or prediction. However, the lack of mining of temporal dependent relationships and diversified event patterns within videos limit the performance of existing methods. To tackle these problems, we propose a novel prototype-guided and dynamic-aware long-distance frame prediction paradigm for VAD. Specifically, we develop a prototype-guided dynamics matching network (PDM-Net) to enhance the discriminant and robustness of anomaly detector. To explore the temporal contexts, we equip PDM-Net with a long short-term dynamic prototype alignment learning mechanism, which stores long-term dynamic prototypes into memory bank and learns how to recall long-term dynamic prototypes with short-term dynamics. As a result, the short input sequences can recall long-term dynamic prototypes stored in the memory bank to achieve the task of long-distance frame prediction. Besides, a feature discrimination module is adopted to extract the representative dynamic features of various normal events meanwhile preserving the diversity of normal patterns. Experimental results on four public datasets demonstrate the superiority of our method.

Binxiao Huang, Jason Chun Lok Li, Jie Ran, Boyu Li, Jiajun Zhou, Dahai Yu, Ngai Wong

Conventional super-resolution (SR) schemes make heavy use of convolutional neural networks (CNNs), which involve intensive multiply-accumulate (MAC) operations, and require specialized hardware such as graphics processing units. This contradicts the regime of edge AI that often runs on devices strained by power, computing, and storage resources. Such a challenge has motivated a series of lookup table (LUT)-based SR schemes that employ simple LUT readout and largely elude CNN computation. Nonetheless, the multi-megabyte LUTs in existing methods still prohibit on-chip storage and necessitate off-chip memory transport. This work tackles this storage hurdle and innovates hundred-kilobyte LUT (HKLUT) models amenable to on-chip cache. Utilizing an asymmetric two-branch multistage network coupled with a suite of specialized kernel patterns, HKLUT demonstrates an uncompromising performance and superior hardware efficiency over existing LUT schemes. Our implementation is publicly available at: https://github.com/jasonli0707/hklut.

Jiacong Hu, Jingwen Ye, Zunlei Feng, Jiazhen Yang, Shunyu Liu, Xiaotian Yu, Lingxiang Jia, Mingli Song

Convolutional Neural Networks (CNNs) are well-known for their vulnerability to adversarial attacks, posing significant security concerns. In response to these threats, various defense methods have emerged to bolster the model's robustness. However, most existing methods either focus on learning from adversarial perturbations, leading to overfitting to the adversarial examples, or aim to eliminate such perturbations during inference, inevitably increasing computational burdens. Conversely, clean training, which strengthens the model's robustness by relying solely on clean examples, can address the aforementioned issues. In this paper, we align with this methodological stream and enhance its generalizability to unknown adversarial examples. This enhancement is achieved by scrutinizing the behavior of latent features within the network. Recognizing that a correct prediction relies on the correctness of the latent feature's pattern, we introduce a novel and effective Feature Pattern Consistency Constraint (FPCC) method to reinforce the latent feature's capacity to maintain the correct feature pattern. Specifically, we propose Spatial-wise Feature Modification and Channel-wise Feature Selection to enhance latent features. Subsequently, we employ the Pattern Consistency Loss to constrain the similarity between the feature pattern of the latent features and the correct feature pattern. Our experiments demonstrate that the FPCC method empowers latent features to uphold correct feature patterns even in the face of adversarial examples, resulting in inherent adversarial robustness surpassing state-of-the-art models.

Haodong Hong, Sen Wang, Zi Huang, Qi Wu, Jiajun Liu

Current Vision-and-Language Navigation (VLN) tasks mainly employ textual instructions to guide agents. However, being inherently abstract, the same textual instruction can be associated with different visual signals, causing severe ambiguity and limiting the transfer of prior knowledge in the vision domain from the user to the agent. To fill this gap, we propose Vision-and-Language Navigation with Multi-modal Prompts (VLN-MP), a novel task augmenting traditional VLN by integrating both natural language and images in instructions. VLN-MP not only maintains backward compatibility by effectively handling text-only prompts but also consistently shows advantages with different quantities and relevance of visual prompts. Possible forms of visual prompts include both exact and similar object images, providing adaptability and versatility in diverse navigation scenarios. To evaluate VLN-MP under a unified framework, we implement a new benchmark that offers: (1) a training-free pipeline to transform textual instructions into multi-modal forms with landmark images; (2) diverse datasets with multi-modal instructions for different downstream tasks; (3) a novel module designed to process various image prompts for seamless integration with state-of-the-art VLN models. Extensive experiments on four VLN benchmarks (R2R, RxR, REVERIE, CVDN) show that incorporating visual prompts would significantly boost navigation performance. While maintaining efficiency with text-only prompts, VLN-MP enables agents to navigate in the pre-explore setting and outperform text-based models, showing its broader applicability. Code is available at https://github.com/honghd16/VLN-MP.

Zheqi He, Xinya Wu, Pengfei Zhou, Richeng Xuan, Guang Liu, Xi Yang, Qiannan Zhu, Hua Huang

Multi-modal large language models(MLLMs) have achieved remarkable progress and demonstrated powerful knowledge comprehension and reasoning abilities. However, the mastery of domain-specific knowledge, which is essential for evaluating the intelligence of MLLMs, continues to be a challenge. Current multi-modal benchmarks for domain-specific knowledge concentrate on multiple-choice questions and are predominantly available in English, which imposes limitations on the comprehensiveness of the evaluation. To this end, we introduce CMMU, a novel benchmark for multi-modal and multi-type question understanding and reasoning in Chinese. CMMU consists of 3,603 questions in 7 subjects, covering knowledge from primary to high school. The questions can be categorized into 3 types: multiple-choice, multiple-response, and fill-in-the-blank, bringing greater challenges to MLLMs. In addition, we propose an evaluation strategy called Positional Error Variance for assessing multiple-choice questions. The strategy aims to perform a quantitative analysis of position bias. We evaluate seven open-source MLLMs along with GPT4-V, Gemini-Pro, and Qwen-VL-Plus. The results demonstrate that CMMU poses a significant challenge to the recent MLLMs. The data and code are available at https://github.com/FlagOpen/CMMU.

Quansong He, Xiaojun Yao, Jun Wu, Zhang Yi, Tao He

In recent years, advanced U-like networks have demonstrated remarkable performance in medical image segmentation tasks. However, their drawbacks, including excessive parameters, high computational complexity, and slow inference speed, pose challenges for practical implementation in scenarios with limited computational resources. Existing lightweight U-like networks have alleviated some problems, but they often have pre-designed structures and consist of non-detachable modules, limiting their application scenarios. In this paper, we propose three plug-and-play decoders by employing different discretization methods of the neural memory Ordinary Differential Equation (nmODE). These decoders integrate features at various levels of abstraction by processing information from skip connections and performing numerical operations on upward paths. Through experiments on the PH2, ISIC2017, and ISIC2018 datasets, we embed these decoders into different U-like networks, demonstrating their effectiveness in significantly reducing the number of parameters and computation while maintaining performance. In summary, the proposed discretized nmODE decoder is capable of reducing the number of parameters by about 20% ~ 50% and computation by up to 74%, while being adaptive to all U-like networks. Our code is available at https://github.com/nayutayuki/Lightweight-nmODE-Decoders-For-U-like-networks.

Qingdong He, Jinlong Peng, Zhengkai Jiang, Kai Wu, Xiaozhong Ji, Jiangning Zhang, Yabiao Wang, Chengjie Wang, Mingang Chen, Yunsheng Wu

3D open-vocabulary scene understanding aims to recognize arbitrary novel categories beyond the base label space. However, existing works not only fail to fully utilize all the available modal information in the 3D domain but also lack sufficient granularity in representing the features of each modality. In this paper, we propose a unified multimodal 3D open-vocabulary scene understanding network, namely UniM-OV3D, aligning point clouds with image, language and depth. To better integrate global and local features of the point clouds, we design a hierarchical point cloud feature extraction module that learns fine-grained feature representations. Further, to facilitate the learning of coarse-to-fine point-semantic representations from captions, we propose the utilization of hierarchical 3D caption pairs, capitalizing on geometric constraints across various viewpoints of 3D scenes. Extensive experimental results have demonstrated the effectiveness and superiority of our method in open-vocabulary semantic and instance segmentation, which achieves state-of-the-art performance on both indoor and outdoor benchmarks such as ScanNet, ScanNet200, S3IDS and nuScenes. Code is available at https://github.com/hithqd/UniM-OV3D.

Jiezhou He, Xue Zhao, Zhiming Luo, Songzhi Su, Shaozi Li, Guojun Zhang

Accurate segmentation of breast tumors in dynamic contrast-enhanced magnetic resonance images (DCE-MRI) is critical for early diagnosis of breast cancer. However, this task remains challenging due to the wide range of tumor sizes, shapes, and appearances. Additionally, the complexity is further compounded by the high dimensionality and ill-posed artifacts present in DCE-MRI data. Furthermore, accurately modeling features in DCE-MRI sequences presents a challenge that hinders the effective representation of essential tumor characteristics. Therefore, this paper introduces a novel Temporal-Spatial Enhanced Network (TSESNet) for breast tumor segmentation in DCE-MRI. TSESNet leverages the spatial and temporal dependencies of DCE-MRI to provide a comprehensive representation of tumor features. To address sequence modeling challenges, we propose a Temporal-Spatial Contrastive Loss (TSCLoss) that maximizes the distance between different classes and minimizes the distance within the same class, thereby improving the separation between tumors and the background. Moreover, we design a novel Temporal Series Feature Fusion (TSFF) module that effectively integrates temporal MRI features from multiple time points, enhancing the model's ability to handle temporal sequences and improving overall performance. Finally, we introduce a simple and effective Tumor-Aware (TA) module that enriches feature representation to accommodate tumors of various sizes. We conducted comprehensive experiments to validate the proposed method and demonstrate its superior performance compared to recent state-of-the-art segmentation methods on two breast cancer DCE-MRI datasets.