International Journal of Computer Vision最新文献_第4页

G2S2-IR: Glocal Gaussian Splatting with Surface Awareness for Inverse Rendering G2S2-IR：具有表面感知的逆渲染的全局高斯溅射

IF 19.5 2区计算机科学

International Journal of Computer Vision Pub Date : 2026-09-28 DOI: 10.1007/s11263-026-03022-8

Junnan Li, Shengping Zhang, Ru Li, Shunyuan Zheng, Chenyang Wang, Henglu Wei, Xiangyang Ji

引用次数: 0

High-Fidelity Implicit Text-to-Shape Generation with Voxelized Diffusion 具有体素化扩散的高保真隐式文本到形状生成

IF 19.5 2区计算机科学

International Journal of Computer Vision Pub Date : 2026-09-27 DOI: 10.1007/s11263-026-03055-z

Yueqi Duan,Fangfu Liu,Hanyang Wang,Muheng Li,Jiwen Lu,Jie Zhou

引用次数: 0

Parameter-Efficient Fine-Tuning for Foundation Models 基础模型的参数高效微调

IF 19.5 2区计算机科学

International Journal of Computer Vision Pub Date : 2026-09-25 DOI: 10.1007/s11263-026-03004-w

Dan Zhang,Tao Feng,Lilong Xue,Yuandong Wang,Yuxiao Dong,Jie Tang

引用次数: 0

TAME-MA: Taming Mobile Agents via Trajectory-Level Trustworthiness Under Jailbreak Manipulations TAME-MA：在越狱操作下通过轨迹级可信度驯服移动代理

IF 19.5 2区计算机科学

International Journal of Computer Vision Pub Date : 2026-09-23 DOI: 10.1007/s11263-026-03015-7

Siyuan Liang,Tianmeng Fang,Yongcheng Jing,Hanqing Zhao,Jiaxing Huang,Dacheng Tao

引用次数: 0

Voxelized 3D Gaussian Representation for Dense Visual SLAM on Embedded Vision System 嵌入式视觉系统中密集视觉SLAM的体素化三维高斯表示

IF 19.5 2区计算机科学

International Journal of Computer Vision Pub Date : 2026-09-23 DOI: 10.1007/s11263-026-03039-z

Tianchen Deng,Chang Nie,Shuhong Liu,Wenhua Wu,Jianfei Yang,Shenghai Yuan,Jiuming Liu,Zhe Liu,Danwei Wang,Hesheng Wang

引用次数: 0

Sampling-Free Universal Domain Adaptive Object Detection with Synergistic Optimization Alignment 基于协同优化对齐的无采样通用域自适应目标检测

IF 19.5 2区计算机科学

International Journal of Computer Vision Pub Date : 2026-09-23 DOI: 10.1007/s11263-026-03047-z

Yuanfan Zheng,Jinlin Wu,Wuyang Li,Kailun Yang,Jun Yu,Zhen Chen

{"title":"Sampling-Free Universal Domain Adaptive Object Detection with Synergistic Optimization Alignment","authors":"Yuanfan Zheng,Jinlin Wu,Wuyang Li,Kailun Yang,Jun Yu,Zhen Chen","doi":"10.1007/s11263-026-03047-z","DOIUrl":"https://doi.org/10.1007/s11263-026-03047-z","url":null,"abstract":"Abstract Universal Domain Adaptive Object Detection (UniDAOD) addresses domain shifts in open-world scenarios without assuming pre-defined shared categories between source and target domains. Despite progress in this field, existing methods rely heavily on thresholding techniques to filter private category samples, which introduces challenges in hyperparameter tuning and handling noisy data. To overcome these limitations, we propose Synergistic Optimization Alignment (SOA), a sampling-free framework that tackles two challenges inherent in UniDAOD: multi-scale discriminator discrepancy and task-dominant gradient conflict. To address the discrepancy among multi-scale discriminators, we propose Multi-Scale Discriminator Alignment (MDA) to adaptively aggregate under-aligned discriminators for shared classes, ensuring consistent feature alignment across scales. To address task-dominant gradient conflicts, we propose Gradient Coordinated Optimization (GCO), which suppresses conflicting domain-alignment gradients and coordinates their magnitudes to prevent domain over-alignment, thereby mitigating private-class interference while preserving beneficial domain-invariant supervision. Extensive experiments on open-set, partial-set, and closed-set domain adaptation benchmarks demonstrate that SOA outperforms state-of-the-art UniDAOD and DAOD methods. Code is available at https://github.com/zyfone/SOA .","PeriodicalId":13752,"journal":{"name":"International Journal of Computer Vision","volume":"16 1","pages":""},"PeriodicalIF":19.5,"publicationDate":"2026-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149049199","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Pseudo-label Refinement and Model-State Recovery for Continual Test-Time Adaptation on Semantic Segmentation 语义分割中连续测试时间自适应的伪标签改进和模型状态恢复

IF 19.5 2区计算机科学

International Journal of Computer Vision Pub Date : 2026-09-23 DOI: 10.1007/s11263-026-03003-x

Kai Hu,Zhihong Zhao,Yuan Zhang,Xieping Gao

引用次数: 0

BD-Diff: Generative Diffusion Model for Image Deblurring on Unknown Domains with Blur-Decoupled Learning 基于模糊解耦学习的未知域图像去模糊生成扩散模型

IF 19.5 2区计算机科学

International Journal of Computer Vision Pub Date : 2026-09-23 DOI: 10.1007/s11263-026-03026-4

Junhao Cheng,Wei-Ting Chen,Xi Lu,Ming-Hsuan Yang

引用次数: 0

PosePAL: Efficient Animal Pose Labeling Using Point Trackers PosePAL：使用点跟踪器的有效动物姿势标记

IF 19.5 2区计算机科学

International Journal of Computer Vision Pub Date : 2026-09-23 DOI: 10.1007/s11263-026-02976-z

Zhuoyang Pan,Boxiao Pan,Guandao Yang,Adam W. Harley,Leonidas Guibas

{"title":"PosePAL: Efficient Animal Pose Labeling Using Point Trackers","authors":"Zhuoyang Pan,Boxiao Pan,Guandao Yang,Adam W. Harley,Leonidas Guibas","doi":"10.1007/s11263-026-02976-z","DOIUrl":"https://doi.org/10.1007/s11263-026-02976-z","url":null,"abstract":"Abstract Automatically estimating animal poses from videos is important for studying animal behavior. Existing methods do not perform reliably, partly because they rely on high-quality training data, and existing such datasets are small. It is challenging and expensive to create such data, due to the large variations in animal morphology and the difficulty of frame-wise video annotation in general. In this paper, we introduce an animal pose labeling pipeline that leverages test-time optimization of a point tracker, providing temporally-consistent annotations at low cost. Given a video, we fine-tune a lightweight appearance embedding inside a pre-trained general-purpose point tracker on a sparse set of annotated frames. These annotations can be obtained from human labelers or off-the-shelf pose detectors. The fine-tuned model is then applied to the rest of the frames for automatic labeling. Our method achieves state-of-the-art performance at a reasonable annotation cost. We believe our pipeline offers a valuable tool for the automatic quantification of animal behavior. Visit our project webpage at https://zhuoyang-pan.github.io/animal-labeling .","PeriodicalId":13752,"journal":{"name":"International Journal of Computer Vision","volume":"52 1","pages":""},"PeriodicalIF":19.5,"publicationDate":"2026-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149049203","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

UniColorFlow: A Unified Framework for Multimodal-Guided Video Colorization with Consistency Flow UniColorFlow：一个统一的框架，用于多模态引导视频着色与一致性流

IF 19.5 2区计算机科学

International Journal of Computer Vision Pub Date : 2026-09-09 DOI: 10.1007/s11263-026-03058-w

Zheng Chang, Lihan Lin, Huan Ouyang, Shuchen Weng, Si Li, Boxin Shi

{"title":"UniColorFlow: A Unified Framework for Multimodal-Guided Video Colorization with Consistency Flow","authors":"Zheng Chang, Lihan Lin, Huan Ouyang, Shuchen Weng, Si Li, Boxin Shi","doi":"10.1007/s11263-026-03058-w","DOIUrl":"https://doi.org/10.1007/s11263-026-03058-w","url":null,"abstract":"Video colorization aims to revitalize grayscale footage with vibrant and temporally consistent colors, yet it remains an inherently ill-posed problem. While user-guided methods offer controllability, existing multimodal frameworks struggle with significant limitations, including limited adherence to conditions due to luminance bias, potential conflicts among guidance signals, inconsistent identities across multi-shot videos, and prohibitive computational costs. To address these challenges, we introduce UniColorFlow, a unified framework for multimodal-guided video colorization. We propose a data augmentation strategy to decouple luminance from color, thereby mitigating inherent biases in the training data. We further design a unified condition network to integrate diverse spatial and language inputs. This module effectively resolves condition conflicts and enforces multi-shot consistency via a shot-aware attention mechanism. Finally, we employ a one-stage training strategy that unifies training and consistency learning, significantly reducing both training and inference overhead. Extensive experiments demonstrate that UniColorFlow outperforms state-of-the-art methods, achieving superior generation quality with flexible controllability and high efficiency.","PeriodicalId":13752,"journal":{"name":"International Journal of Computer Vision","volume":"22 1","pages":""},"PeriodicalIF":19.5,"publicationDate":"2026-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149086485","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0