IEEE Transactions on Circuits and Systems for Video Technology最新文献

{"title":"Guided LDR Generation for Single-Image HDR Reconstruction Using Exposure Priors From Multi-Exposure Images","authors":"Bolun Zheng;Shun Guo;Qianyu Zhang;Han Wang;Quan Chen;Tao Zhang;Ying Fu","doi":"10.1109/TCSVT.2026.3709976","DOIUrl":"https://doi.org/10.1109/TCSVT.2026.3709976","url":null,"abstract":"Single-image high dynamic range (HDR) reconstruction remains challenging due to exposure-induced degradations. Several recent methods attempt to synthesize multi-exposure low dynamic range (LDR) images from a single input and fuse them to recover missing content. However, they primarily rely on implicit statistical priors learned from data and do not explicitly extract or constrain exposure-related degradation patterns, which limits their ability to accurately restore detail and color under extreme lighting conditions. To address this, we present a two-stage framework, Degradation-Guided Single-Image HDR (DG-SHDR), which formulates HDR reconstruction as a pipeline of degradation modeling and degradation-aware image generation. Our framework consists of two core components: 1) a degradation estimation network (DEN) that learns the image exposure prior from medium-low and medium-high exposure image pairs. 2) a U-shaped degradation-guided reconstruction network (DGR) intended to integrate the exposure prior and restore more accurate low- and high-exposure images. These generated images are subsequently fused to produce a high-quality HDR output. Extensive quantitative and qualitative experimental results demonstrate that our method excels in detail, brightness, and color restoration, significantly improving the overall performance of single-image HDR reconstruction. The codes and models will be available at <uri>https://github.com/gguoshun/DG-SHDR</uri>","PeriodicalId":13082,"journal":{"name":"IEEE Transactions on Circuits and Systems for Video Technology","volume":"36 10","pages":"15478-15491"},"PeriodicalIF":10.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097495","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dynamic Gaussian Splatting With Efficient Depth Optimization and Regularization for Sparse-View Endoscopic Scene Reconstruction","authors":"Hengjia Hu;Nanqing Dong;Mounir Kaaniche;Congcong Wang;Wanli Ouyang;Shengyong Chen","doi":"10.1109/TCSVT.2026.3710337","DOIUrl":"https://doi.org/10.1109/TCSVT.2026.3710337","url":null,"abstract":"In endoscopic surgery, reconstructing dynamic scenes using 4D Gaussian Splatting (4DGS) offers high-quality results but suffers from severe geometric instability under sparse-view conditions. This instability stems from geometrically inconsistent priors and the lack of a synergistic regularization scheme. To address these root causes, we introduce a novel framework built upon two complementary modules. First, a <italic>Depth Optimization</i> framework produces a scale-consistent geometric prior by fusing multi-resolution depth maps and anchoring the scale via sparse structure-from-motion (SfM) points. Second, a <italic>Multi-faceted Geometric Regularization</i> scheme employs a trio of complementary constraints (across global-local depth, virtual views, and surface normals) that work in synergy to mitigate severe overfitting. Extensive experiments on two public surgical datasets demonstrate superior sparse-view reconstruction performance over representative baselines at real-time rendering speeds. This work delivers a robust and efficient solution for high-fidelity 4D scene reconstruction in minimally invasive endoscopy.","PeriodicalId":13082,"journal":{"name":"IEEE Transactions on Circuits and Systems for Video Technology","volume":"36 10","pages":"14800-14812"},"PeriodicalIF":10.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097450","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"IEEE Transactions on Circuits and Systems for Video Technology Publication Information","authors":"","doi":"10.1109/TCSVT.2026.3735812","DOIUrl":"https://doi.org/10.1109/TCSVT.2026.3735812","url":null,"abstract":"","PeriodicalId":13082,"journal":{"name":"IEEE Transactions on Circuits and Systems for Video Technology","volume":"36 10","pages":"C2-C2"},"PeriodicalIF":10.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11720476","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097723","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"One-Shot Unsupervised Domain Adaptation for Semantic Segmentation With Self-Constrained Style Decoupling","authors":"Jinghui Qin;Hao Tan;Kebing Jin","doi":"10.1109/TCSVT.2026.3702681","DOIUrl":"https://doi.org/10.1109/TCSVT.2026.3702681","url":null,"abstract":"One-shot unsupervised domain adaptation (OSUDA) aims to adapt a segmentation model from a labeled source domain to a target domain where only a single unlabeled datum is available. Although style transfer-based techniques, where the source images are stylized to have the appearance of the target domain, have shown impressive performance, they often suffer from a lack of content diversity, style distortion, or style overfitting. To mitigate these issues, we propose a novel Self-Constrained Style Decoupling (SCSD) framework to directly decouple the style of an image, extend the target style space, and leverage a text-to-image diffusion model to generate a synthetic target dataset with diverse photo-realistic images following the style of the target domain. Specifically, to realize OSUDA for semantic segmentation, SCSD carries out three learning stages, including a style decoupling stage, a self-constraining fine-tuning stage, and an adaptive segmentation stage. The style vector of the single target image is first decoupled by a style decoupling network in the style decoupling stage. Then, a fine-tuned text-to-image diffusion model with the style vectors of the target image’s patch-shuffled variants is deployed to generate style-diversified images aligned with the target style space. At the final adaptive segmentation stage, we integrate the source domain and generated target domain datasets to train the generalized UDA segmentation model. Our SCSD can be seamlessly integrated with any UDA segmentation models, underscoring the practical value. Extensive experiments on the Cityscapes, GTA, and SYNTHIA datasets show that our SCSD can achieve superior performance by outperforming state-of-the-art OSUDA and T2I personalization methods in the realistic synthetic source domain to real-world target domain adaptation task under the one-shot setting.","PeriodicalId":13082,"journal":{"name":"IEEE Transactions on Circuits and Systems for Video Technology","volume":"36 10","pages":"15864-15876"},"PeriodicalIF":10.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097851","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"CASS: A Context-Aware Memory Framework With Spectral–Spatial Modulation for UAV Multispectral Video Tracking","authors":"Yuzeng Chen;Qiangqiang Yuan;Bineng Zhong;Xiantao Hu;Jie Zhou;Xi Zhao;Xin Su;Yuqi Tang;Liangpei Zhang","doi":"10.1109/TCSVT.2026.3708187","DOIUrl":"https://doi.org/10.1109/TCSVT.2026.3708187","url":null,"abstract":"Unmanned Aerial Vehicle (UAV) multispectral video object tracking is critical for real-world applications. While multispectral imaging offers complementary spectral cues beyond the visible range, tracking in aerial scenarios remains challenging due to data scarcity, suboptimal spectral-spatial modulation, and discrete sequential temporal modeling. To this end, we propose CASS, a context-aware memory framework with spectral-spatial modulation, which integrates spectral, spatial, and temporal cues for UAV multispectral tracking. CASS introduces two lightweight modules: 1) the efficient spectral-spatial modulation (ESSM) module, which modulates spatial representations through spectral-guided fusion, and 2) the context state space reasoning (CSSR) module, which leverages evolving state space memory to retain long-term temporal cues and mitigate error propagation during cross-frame reasoning. By integrating these components in a parameter-efficient fine-tuning fashion, CASS achieves both efficient modulation and context-aware tracking. Evaluations on the UAV multispectral benchmark (MUST) and ground-based hyperspectral benchmarks (NIR, RedNIR, VIS, MSSOT, MSVT) demonstrate CASS’s superior performance for both general hyperspectral tracking and specific UAV perception.","PeriodicalId":13082,"journal":{"name":"IEEE Transactions on Circuits and Systems for Video Technology","volume":"36 10","pages":"14681-14695"},"PeriodicalIF":10.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097038","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MonoHPE: Monocular 3D Object Detection With Height Positional Encoding","authors":"Chenxi Zhu;Zhi Li;Zhenyu Dai;Long Zheng;Weidong Wang","doi":"10.1109/TCSVT.2026.3700175","DOIUrl":"https://doi.org/10.1109/TCSVT.2026.3700175","url":null,"abstract":"Monocular 3D object detection in autonomous driving has achieved revolutionary breakthroughs, primarily attributed to integrating perspective projection and geometric priors. Despite the continuous surge in overall performance, we argue that the subtle yet critical challenge of insufficient geometric constraints in the height dimension has not been sufficiently dissected, resulting in a systematic deviation between the bounding box height and the actual projected height, thereby nonlinearly amplifying the error in depth estimation. To uncover the profound essence of this error amplification mechanism, this paper employs a differentiable analytical framework to dissect the ratio dynamics of depth perception in perspective projection, providing a principled explanation of the sensitivity of depth perception to height deviations. Inspired by these insights, we propose MonoHPE, a method based on a hierarchical task learning paradigm. This method explicitly strengthens the geometric associations in the height dimension through an ingenious position modeling mechanism, achieving a breakthrough leap in the precision of depth estimation. Furthermore, to enhance the model’s adaptability in complex environments, we innovatively introduce an adaptive activation-guided loss, which directs the network to focus on semantically dominant regions, effectively suppressing foreground interference and noise contamination. Experimental results demonstrate that our proposed method achieves state-of-the-art performance on both the KITTI and Waymo datasets.","PeriodicalId":13082,"journal":{"name":"IEEE Transactions on Circuits and Systems for Video Technology","volume":"36 10","pages":"14334-14346"},"PeriodicalIF":10.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097190","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DIMQ: Distribution-Induced Multi-Peak Quantization for Deep Neural Networks","authors":"Chao Feng;Guoxu Zhou;Hongyuan Lin;Yi Lin;Xun Wu;Qibin Zhao;Shengli Xie","doi":"10.1109/TCSVT.2026.3707561","DOIUrl":"https://doi.org/10.1109/TCSVT.2026.3707561","url":null,"abstract":"Model quantization is critical for deploying deep neural networks on resource constrained devices. However, conventional quantization schemes typically rely on passive adaptation to inherent weight distributions, resulting in significant quantization noise and codebook collapse due to the structural mismatch with discrete levels. To address this limitation, we propose a weight-side Distribution-Induced Multi-peak Quantization (DIMQ) framework that actively reshapes weight distributions to align with the quantizer topology. DIMQ employs an entropy-regularized soft assignment to progressively organize weights into compact and well-separated peaks, together with a separation constraint that prevents codebook collapse under limited bit budgets. The resulting objective integrates task loss with a differentiable distortion surrogate, facilitating the simultaneous and stable optimization of quantization centers and weight structures via closed-form Gibbs assignment and gradient-stabilized center optimization. Theoretically, we derive rigorous distortion bounds proving that the induced multi-peak distribution yields strictly lower quantization error than gaussian and refined uniform baselines under specific variance conditions. Extensive experiments on ImageNet and COCO benchmarks demonstrate that DIMQ significantly outperforms baselines on CNN architectures and achieves highly competitive accuracy on Transformer models, particularly demonstrating superior robustness in extremely low bit regimes.","PeriodicalId":13082,"journal":{"name":"IEEE Transactions on Circuits and Systems for Video Technology","volume":"36 10","pages":"15946-15961"},"PeriodicalIF":10.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097192","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Motion Saliency Guided Fine-Grained Multimodal Interactive Learning for RGB-D Action Recognition","authors":"Hanbo Wu;Chaoqun Wang;Yimeng Duan;Rui Song","doi":"10.1109/TCSVT.2026.3708969","DOIUrl":"https://doi.org/10.1109/TCSVT.2026.3708969","url":null,"abstract":"RGB-D action recognition remains challenging due to the difficulty of effectively integrating complementary RGB and depth information. Most existing methods rely on feature-level fusion after independent encoding, limiting fine-grained cross-modal interaction. CLIP has recently been adapted for video tasks thanks to its powerful vision-language pretraining, but it typically requires additional temporal modules and has not been applied to RGB-D multimodal fusion for action recognition. In this paper, we innovatively extend CLIP for RGB-D action recognition, and propose a novel motion saliency guided fine-grained multimodal interactive learning (MS-FMIL) framework. Specifically, the RGB image and its corresponding depth residual dynamic image at each time step are jointly fed into CLIP image encoder, enabling patch-level fine-grained interaction between static appearance and dynamic motion cues. We further propose a motion-salient patch selection mechanism, where patch-wise motion saliency scores are computed from depth residual dynamic image to select informative regions across modalities. This enhances interaction efficiency and discriminability while reducing computation redundancy. Our method also enables temporal-aware representation learning without explicit temporal modeling modules, by injecting local spatiotemporal context from depth residual dynamic images directly into CLIP. Moreover, it supports weak temporal alignment, allowing depth dynamic information to complement static RGB features without strict frame-level synchronization. Extensive experiments on benchmark datasets demonstrate the effectiveness of our approach, offering a new perspective on leveraging CLIP for multimodal action recognition.","PeriodicalId":13082,"journal":{"name":"IEEE Transactions on Circuits and Systems for Video Technology","volume":"36 10","pages":"14771-14785"},"PeriodicalIF":10.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097396","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dual Sparse Aggregation Transformer for Multispectral Object Detection","authors":"Wencong Wu;Xiuwei Zhang;Hanlin Yin;Hongxi Zhang;Yanning Zhang","doi":"10.1109/TCSVT.2026.3709222","DOIUrl":"https://doi.org/10.1109/TCSVT.2026.3709222","url":null,"abstract":"Transformer-based approaches have obtained excellent performance in multispectral object detection tasks due to their ability to model long-range dependencies and capture complementary information. However, previous transformer-based multispectral detection methods tend to use all available tokens for similarity calculation, which results in redundant information interaction from irrelevant areas, leading to degraded detection performance. To overcome this challenge, we propose a novel Dual Sparse Aggregation Transformer (DSAFormer) for multispectral object detection, which consists of a Dual Sparse Transformer (DSFormer) and a Learnable Addition Fusion Block (LAFB). Specifically, the DSFormer is designed to exploit and boost cross-modal complementary information, thereby improving detection performance. It incorporates three key components: A Spatial Sparse Multi-Head Cross-Attention (SSMHCA) mechanism selectively captures cross-modal relationships at the spatial level by reserving only the high query-key similarity scores, eliminating irrelevant interactions. A Channel Sparse Multi-Head Cross-Attention (CSMHCA) mechanism performs similar sparse calculations at the channel level to enhance feature representation and filter out low matching query-key. A Multi-Scale Feature Refinement Layer (MSFRL) is developed to aggregate hierarchical features and suppress redundant information. To effectively fuse multimodal features, the LAFB is introduced to aggregate intramodal and intermodal feature information by feature reweighting. Extensive experimental results have demonstrated that our proposed DSAFormer achieves better detection performance against state-of-the-art methods on four public datasets, including the MFAD, FLIR, M³FD, and LLVIP. The source code of our DSAFormer will be released at <uri>https://github.com/WenCongWu/DSAFormer</uri>","PeriodicalId":13082,"journal":{"name":"IEEE Transactions on Circuits and Systems for Video Technology","volume":"36 10","pages":"14549-14564"},"PeriodicalIF":10.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097540","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ICMIL: An Interactive Contrastive Multiple Instance Learning Framework Using Foundation Model Features for Histopathology Image Classification","authors":"Jingjiao Lou;Qingtao Pan;Qing Yang;Guang Yang;Bing Ji;Shuo Li","doi":"10.1109/TCSVT.2026.3708293","DOIUrl":"https://doi.org/10.1109/TCSVT.2026.3708293","url":null,"abstract":"Multiple instance learning (MIL) is a crucial paradigm addressing weakly supervised classification in histopathological images. However, existing MIL methods struggle to model tile/patch interactions, which can capture important contextual information. Moreover, MIL is limited by suboptimal bag embeddings, as traditional methods focus primarily on extracting distinct embeddings for individual instances rather than for the entire bag. These limitations restrict MIL’s ability to leverage contextual information and generate discriminative aggregated representations. To address these limitations, we propose Interactive Contrastive Multiple Instance Learning (ICMIL), a novel MIL framework that integrates graph learning (GL) and contrastive learning (CL) to enhance MIL’s contextual awareness and discriminative power. ICMIL introduces two key innovations: 1) Transformer-based Graph Attention (TransGAT) models comprehensive patch interactions by constructing fully connected graphs and generating long-range edge attention for information aggregation. This approach addresses the problem of limited patch interactions in MIL and boosts MIL’s contextual awareness. 2) Reinforced Contrastive MIL (ReCMIL) refines the bag embedding space by selecting the most advantageous contrastive bag pairs using a policy network. ReCMIL addresses the problem of suboptimal bag embeddings, enabling MIL to generate discriminative aggregated embeddings. Experimental results demonstrate the superiority of our proposed ICMIL method over state-of-the-art approaches on four publicly available datasets, which span three anatomical sites and encompass five classification tasks (including both binary and multiclass classification). Furthermore, we extend ICMIL with features from a pretrained foundation model and achieve the best performance. Specifically, ICMIL achieves validation accuracies of 92.39% on CRC-DX, 86.86% on CRC-KR, 100.00% on BRACS (binary), 66.67% on BRACS (multiclass), and 96.04% on TCGA-Lung. These findings highlight the strong potential of integrating ICMIL with pretrained foundation models for histopathology image analysis. The code will be available at: <uri>https://github.com/JingjiaoLou/ICMIL</uri>","PeriodicalId":13082,"journal":{"name":"IEEE Transactions on Circuits and Systems for Video Technology","volume":"36 10","pages":"15187-15205"},"PeriodicalIF":10.8,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149097452","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}