arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~

期刊&会议

International Conference on Computer Vision · 会议 · Computer Vision

至 收录 4770
2601.13578 2026-01-21 cs.LG cs.CV

FG-OrIU: Towards Better Forgetting via Feature-Gradient Orthogonality for Incremental Unlearning

FG-OrIU: 通过特征-梯度正交性实现更好的增量遗忘

Qian Feng, JiaHang Tu, Mintong Kang, Hanbin Zhao, Chao Zhang, Hui Qian

机构 * College of Computer Science and Technology, Zhejiang University(浙江大学计算机科学与技术学院)

AI总结 FG-OrIU通过特征和梯度正交约束实现深度不可逆的增量遗忘,解决现有方法中残留信息可恢复的问题。

Comments This paper has been accepted by ICCV 2025. code: \url{https://github.com/RAIAN08/FG-OrIU}

URL PDF HTML 收藏
2601.12671 2026-01-21 cs.CV cs.AI cs.LG

Exploiting Test-Time Augmentation in Federated Learning for Brain Tumor MRI Classification

在联邦学习中利用测试时增强进行脑肿瘤MRI分类

Thamara Leandra de Deus Melo, Rodrigo Moreira, Larissa Ferreira Rodrigues Moreira, André Ricardo Backes

机构 * Institute of Exact and Technological Sciences, Federal University of Viçosa - UFV, Rio Paranaíba-MG, Brazil(精确与技术科学研究所,弗拉维亚联邦大学-UFV,里奥帕拉纳伊巴-MG,巴西) Department of Computing, Federal University of São Carlos, São Carlos-SP, Brazil(计算系,萨o卡洛斯联邦大学,萨o卡洛斯-SP,巴西)

AI总结 本文提出在联邦学习中结合测试时增强和轻量预处理以提升脑肿瘤MRI分类的准确性。

Comments 21st International Conference on Computer Vision Theory and Applications (VISAPP 2026), 9-11 March 2026, Marbella, Spain

URL PDF HTML 收藏
2601.12664 2026-01-21 cs.CV cs.AI cs.LG

Generalizable Hyperparameter Optimization for Federated Learning on Non-IID Cancer Images

可泛化的联邦学习超参数优化用于非独立同分布癌症图像

Elisa Gonçalves Ribeiro, Rodrigo Moreira, Larissa Ferreira Rodrigues Moreira, André Ricardo Backes

机构 * Institute of Exact and Technological Sciences, Federal University of Viçosa - UFV, Rio Paranaíba-MG, Brazil(精确与技术科学研究所,弗拉维亚联邦大学-UFV,里奥帕拉纳伊巴-巴西) Department of Computing, Federal University of São Carlos, São Carlos-SP, Brazil(计算系,圣卡洛斯联邦大学,圣卡洛斯-巴西)

AI总结 本文提出了一种简单的方法,通过跨数据集聚合优化器和批量大小,实现了在非IID联邦学习场景中具有竞争力的分类性能。

Comments 21st International Conference on Computer Vision Theory and Applications (VISAPP 2026), 9-11 March 2026, Marbella, Spain

URL PDF HTML 收藏
2601.11635 2026-01-21 cs.CV

Now You See Me, Now You Don't: A Unified Framework for Expression Consistent Anonymization in Talking Head Videos

现在你看见了,现在你又看不见:一种用于说话头视频中表达一致匿名化的统一框架

Anil Egin, Andrea Tangherloni, Antitza Dantcheva

AI总结 Anon-NET通过基于扩散的生成模型和视频驱动动画,实现说话头视频中表达一致的匿名化,同时保留视觉真实性和时间一致性。

Journal ref Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, IEEE/CVF, Oct 2025, Hawaii-Honolulu, United States. pp.5925-5934

URL PDF HTML 收藏
2308.06712 2026-01-21 cs.CV

Compositional Feature Augmentation for Unbiased Scene Graph Generation

组合特征增强用于无偏场景图生成

Lin Li, Guikun Chen, Jun Xiao, Yi Yang, Chunping Wang, Long Chen

机构 * Zhejiang University(浙江大学) The Hong Kong University of Science and Technology(香港理工大学) FinVolution

AI总结 本文提出组合特征增强策略,通过增强关系三元组特征多样性,解决SGG中的偏见问题。

Comments ICCV

URL PDF HTML 收藏
2601.08371 2026-01-14 cs.CV cs.AI cs.GR cs.LG

Geo-NVS-w: Geometry-Aware Novel View Synthesis In-the-Wild with an SDF Renderer

Geo-NVS-w:基于SDF渲染器的野外几何感知新视角合成

Anastasios Tsalakopoulos, Angelos Kanlis, Evangelos Chatzis, Antonis Karakottas, Dimitrios Zarpalas

机构 * Centre for Research and Technology Hellas (CERTH)(希腊研究中心与技术中心)

AI总结 Geo-NVS-w通过SDF渲染器实现野外高保真新视角合成,采用几何保持损失提升细节精度,显著降低能耗。

Comments Presented at the ICCV 2025 Workshop on Large Scale Cross Device Localization

URL PDF HTML 收藏
2509.06387 2026-01-14 cs.CV

Your Super Resolution Model is not Enough for Tackling Real-World Scenarios

你的超分辨率模型不足以应对真实场景

Dongsik Yoon, Jongeun Kim

机构 * HDC LABS(HDC实验室)

AI总结 本文提出SAAM模块,通过轻量级特征提取和注意力机制,提升超分辨率模型在多种缩放因子下的泛化能力,实现高效且鲁棒的多尺度上采样。

Comments To appear in the Workshop on Efficient Computing under Limited Resources: Visual Computing (ECLR) at ICCV 2025

URL PDF HTML 收藏
2601.06883 2026-01-13 cs.CV cs.AI

MixRI: Mixing Features of Reference Images for Novel Object Pose Estimation

MixRI:用于新颖物体姿态估计的参考图像特征混合

Xinhang Liu, Jiawei Shi, Zheng Dang, Yuchao Dai

机构 * School of Electronics and Information, Northwestern Polytechnical University(电子与信息学院,西北工业大学) Shaanxi Key Laboratory of Information Acquisition and Processing(陕西省信息获取与处理重点实验室) CVLab, EPFL, Switzerland(EPFL瑞士计算机视觉实验室)

AI总结 MixRI通过轻量级网络和参考图像融合策略,实现新颖物体姿态估计,减少计算和存储需求,同时保持与传统方法相当的性能。

Comments Accepted by ICCV 2025

Journal ref Proceedings of the IEEE/CVF International Conference on Computer Vision (2025) 9024--9035

URL PDF HTML 收藏
2503.15986 2026-01-12 cs.NE cs.CV

SpiLiFormer: Enhancing Spiking Transformers with Lateral Inhibition

SpiLiFormer:通过横向抑制增强脉冲变换器

Zeqi Zheng, Yanchen Huang, Yingchao Yu, Zizheng Zhu, Junfeng Tang, Zhaofei Yu, Yaochu Jin

机构 * Zhejiang University(浙江大学) Westlake University(西湖大学) Nanjing University(南京大学) Donghua University(东华大学) University of Electronic Science and Technology of China(电子科技大学) Peking University(北京大学)

AI总结 SpiLiFormer通过引入横向抑制机制,提升脉冲变换器对相关上下文的注意力,从而在多个数据集上取得更优性能。

Comments Accepted by ICCV 2025. The first two authors contributed equally

Journal ref Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2025, pp. 24539-24548

URL PDF HTML 收藏
2503.09527 2026-01-12 cs.CV cs.AI

CombatVLA: An Efficient Vision-Language-Action Model for Combat Tasks in 3D Action Role-Playing Games

CombatVLA: 一种高效的视觉-语言-动作模型,用于3D动作角色扮演游戏中的战斗任务

Peng Chen, Pi Bu, Yingyao Wang, Xinyi Wang, Ziming Wang, Jie Guo, Yingxiu Zhao, Qi Zhu, Jun Song, Siran Yang, Jiamang Wang, Bo Zheng

机构 * Alibaba Group(阿里巴巴集团)

AI总结 CombatVLA是一种高效的视觉-语言-动作模型,用于3D动作角色扮演游戏中的战斗任务,通过高效的训练和推理方法,在战斗理解基准上表现优异,并实现了游戏战斗的50倍加速。

Comments Accepted by ICCV 2025

URL PDF HTML 收藏
2601.04968 2026-01-09 cs.CV

SparseLaneSTP: Leveraging Spatio-Temporal Priors with Sparse Transformers for 3D Lane Detection

SparseLaneSTP: 利用稀疏变换器结合时空先验进行3D车道检测

Maximilian Pittner, Joel Janai, Mario Faigle, Alexandru Paul Condurache

机构 * Bosch Mobility Solutions, Robert Bosch GmbH(博世移动解决方案,罗伯特·博世有限公司) Institute of Neuro- and Bioinformatics, University of Lübeck(神经与生物医学研究所,吕贝克大学) Institute for Signal Processing and System Theory, University of Stuttgart(信号处理与系统理论研究所,斯图加特大学)

AI总结 SparseLaneSTP通过整合车道结构的几何属性和时间信息,利用稀疏变换器和时空注意力机制,提升3D车道检测的精度和一致性。

Comments Published at IEEE/CVF International Conference on Computer Vision (ICCV) 2025

URL PDF HTML 收藏
2601.03884 2026-01-08 cs.CV cs.AI

FLNet: Flood-Induced Agriculture Damage Assessment using Super Resolution of Satellite Images

FLNet:利用卫星图像超分辨率进行洪水引起的农业损害评估

Sanidhya Ghosal, Anurag Sharma, Sushil Ghildiyal, Mukesh Saini

机构 * Annam.AI CoE, MoE, Indian Institute of Technology Ropar, Punjab, India(Annam.AI联合研究所、教育部、印度理工学院罗帕尔分校) Department of Mathematical Sciences, Rajiv Gandhi Institute of Petroleum Technology, Jais, UP, India(数学科学系、拉贾吉安石油技术学院) Department of Computer Science and Engineering, Indian Institute of Technology Ropar, Punjab, India(计算机科学与工程系、印度理工学院罗帕尔分校)

AI总结 FLNet利用卫星图像超分辨率技术,提升空间分辨率以更准确评估洪水导致的农业损害,实现高效、低成本的自动化评估。

Comments Accepted for oral presentation at the 10th International Conference on Computer Vision and Image Processing (CVIP 2025)

URL PDF HTML 收藏
2601.03392 2026-01-08 cs.CV cs.NE

Better, But Not Sufficient: Testing Video ANNs Against Macaque IT Dynamics

更强大,但不足够:测试视频ANNs against Macaque IT动态

Matteo Dunnhofer, Christian Micheloni, Kohitij Kar

机构 * York University(约克大学) University of Udine(乌迪内大学)

AI总结 本文通过比较自然视频与静态、递归和视频基于的ANN模型,发现视频模型在神经预测性上有所改进,但无法捕捉IT中表达的外观不变的时间计算,表明需要新的目标来编码生物时间统计和不变性。

Comments Extended Abstract at the 2nd Human-inspired Computer Vision workshop at ICCV 2025

URL PDF HTML 收藏
2504.07960 2026-01-08 cs.CV

VisualCloze: A Universal Image Generation Framework via Visual In-Context Learning

VisualCloze: 一种通过视觉上下文学习实现通用图像生成的框架

Zhong-Yu Li, Ruoyi Du, Juncheng Yan, Le Zhuo, Qilong Wu, Zhen Li, Peng Gao, Zhanyu Ma, Ming-Ming Cheng

机构 * VCIP, CS, Nankai University(VCIP、计算机科学、南开大学) Beijing University of Posts and Telecommunications(北京邮电大学) Tsinghua University(清华大学) Shanghai AI Laboratory(上海人工智能实验室) The Chinese University of Hong Kong(香港中文大学)

AI总结 VisualCloze通过视觉上下文学习实现通用图像生成,支持多种任务泛化与反向生成,利用图结构数据集提升任务密度和知识迁移。

Comments Accepted at ICCV 2025. Project page: https://visualcloze.github.io

URL PDF HTML 收藏
2601.02759 2026-01-07 cs.CV cs.RO

Towards Zero-Shot Point Cloud Registration Across Diverse Scales, Scenes, and Sensor Setups

跨多样尺度、场景和传感器设置的零样本点云配准

Hyungtae Lim, Minkyun Seo, Luca Carlone, Jaesik Park

机构 * Laboratory for Information and Decision Systems (LIDS), Massachusetts Institute of Technology(信息与决策系统实验室,麻省理工学院) Department of Computer Science and Engineering, Seoul National University(计算机科学与工程系,首尔国立大学)

AI总结 提出BUFFER-X框架,通过几何自举、分布感知采样和坐标归一化实现零样本点云配准,同时引入BUFFER-X-Lite提升效率,适用于多样场景和传感器设置。

Comments 18 pages, 15 figures. Extended version of our ICCV 2025 highlight paper [arXiv:2503.07940]. arXiv admin note: substantial text overlap with arXiv:2503.07940

URL PDF HTML 收藏
2503.06472 2026-01-07 cs.CV cs.MM

CalliReader: Contextualizing Chinese Calligraphy via an Embedding-Aligned Vision-Language Model

CalliReader:通过嵌入对齐的视觉-语言模型 contextualizing 中文书法

Yuxuan Luo, Jiaqi Tang, Chenyi Huang, Feiyang Hao, Zhouhui Lian

机构 * Wangxuan Institute of Computer Technology, Peking University(北京大学计算机技术研究院)

AI总结 CalliReader通过字符级切片、视觉-文本对齐和嵌入指令微调,解决中文书法的上下文识别问题,实现比现有方法和专业书法家更高的准确性与更低的幻觉率。

Comments 11 pages

Journal ref Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) 2025

URL PDF HTML 收藏
2601.02339 2026-01-06 cs.CV

Joint Semantic and Rendering Enhancements in 3D Gaussian Modeling with Anisotropic Local Encoding

在3D高斯建模中结合语义和渲染增强的各向异性局部编码

Jingming He, Chongyi Li, Shiqi Wang, Sam Kwong

机构 * City University of Hong Kong(香港城市大学) Nankai University(南开大学) Lingnan University(岭南大学)

AI总结 本文提出一种联合增强框架,通过各向异性局部编码提升3D高斯建模的语义和渲染性能,实现更精确的形状表示和高效的渲染效果。

Comments Accepted by ICCV 2025

URL PDF HTML 收藏
2411.17984 2026-01-06 cs.CV

RS-vHeat: Heat Conduction Guided Efficient Remote Sensing Foundation Model

RS-vHeat:基于热传导的高效远程传感基础模型

Huiyang Hu, Peijin Wang, Hanbo Bi, Boyuan Tong, Zhaozhi Wang, Wenhui Diao, Hao Chang, Yingchao Feng, Ziqi Zhang, Yaowei Wang, Qixiang Ye, Kun Fu, Xian Sun

AI总结 RS-vHeat通过热传导运算符和自监督策略,实现高效多模态遥感基础模型,提升效率和性能。

Comments 19 pages, 8 figures and 10 tables

Journal ref Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2025

URL PDF HTML 收藏
2309.07926 2026-01-06 eess.IV cs.CV

COMPASS: High-Efficiency Deep Image Compression with Arbitrary-scale Spatial Scalability

COMPASS:基于深度学习的高效率图像压缩与任意尺度空间可扩展性

Jongmin Park, Jooyoung Lee, Munchurl Kim

机构 * KAIST(韩国科学技术院) ETRI(电子与信息技术研究所)

AI总结 COMPASS是一种基于深度学习的高效率图像压缩方法,支持任意尺度的空间可扩展性,通过LIFF方法和联合RD损失函数实现高效编码。

Comments Accepted in ICCV 2023. Please visit our project page at https://kaist-viclab.github.io/compass-site/

URL PDF HTML 收藏
2512.24097 2026-01-01 cs.CV cs.AI cs.CL cs.MM

Factorized Learning for Temporally Grounded Video-Language Models

分解学习用于时间感知的视频-语言模型

Wenzheng Zeng, Difei Gao, Mike Zheng Shou, Hwee Tou Ng

机构 * National University of Singapore(新加坡国立大学)

AI总结 本文提出D$^2$VLM框架,通过分解学习方法提升视频-语言模型在时间定位和文本响应任务中的性能,引入证据标记和FPO算法以优化学习过程。

Comments ICCV 2025 paper. This arXiv version updates Figure 1 to include the concurrent work Qwen2.5-VL to ensure consistency with Table 1

URL PDF HTML 收藏
2507.22731 2026-01-01 cs.MM

GestureHYDRA: Semantic Co-speech Gesture Synthesis via Hybrid Modality Diffusion Transformer and Cascaded-Synchronized Retrieval-Augmented Generation

GestureHYDRA: 通过混合模态扩散变换器和级联同步检索增强生成进行语义同步手势合成

Quanwei Yang, Luying Huang, Kaisiyuan Wang, Jiazhi Guan, Shengyi He, Fengguo Li, Hang Zhou, Lingyun Yu, Yingying Li, Haocheng Feng, Hongtao Xie

AI总结 GestureHYDRA通过混合模态扩散变换器和级联同步检索增强生成技术,实现具有明确语义的手势合成,提升手势生成的准确性和效率。

Comments 10 pages, 5 figures, Accepted by ICCV 2025

URL PDF HTML 收藏
2504.21414 2026-01-01 cs.CV

Adapting In-Domain Few-Shot Segmentation to New Domains without Source Domain Retraining

在新领域中无需源领域再训练即可适应领域内少样本分割

Qi Fan, Kaiqi Liu, Nian Liu, Hisham Cholakkal, Rao Muhammad Anwer, Wenbin Li, Yang Gao

机构 * Nanjing University(南京大学) Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学)

AI总结 本文提出ISA方法,通过在推理过程中学习领域特征,适应已训练的FSS模型结构,无需源领域再训练,提升跨领域少样本分割性能。

Comments Accepted by ICCV 2025

URL PDF HTML 收藏
2503.09131 2025-12-30 cs.CV

MP-HSIR: A Multi-Prompt Framework for Universal Hyperspectral Image Restoration

MP-HSIR: 一种多提示框架用于通用超光谱图像修复

Zhehui Wu, Yong Chen, Naoto Yokoya, Wei He

机构 * Wuhan University(武汉大学) Jiangxi Normal University(江西师范大学) The University of Tokyo(东京大学) RIKEN Center for Advanced Intelligence Project(理化学研究所先进智能项目中心)

AI总结 MP-HSIR通过多提示框架实现通用超光谱图像修复,整合光谱、文本和视觉提示,提升修复效果和泛化能力。

Comments Accepted to the IEEE/CVF International Conference on Computer Vision (ICCV) 2025

Journal ref Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) 2025, pages 13009-13020

URL PDF HTML 收藏
2503.05332 2025-12-29 cs.CV

CoMoGaussian: Continuous Motion-Aware Gaussian Splatting from Motion-Blurred Images

CoMoGaussian: 从运动模糊图像中连续的运动感知高斯点云生成

Jungho Lee, Donghyeong Kim, Dogyoon Lee, Suhwan Cho, Minhyeok Lee, Wonjoon Lee, Taeoh Kim, Dongyoon Wee, Sangyoun Lee

机构 * Yonsei University(延世大学) NAVER Cloud(NAVER云)

AI总结 CoMoGaussian通过连续运动感知的高斯点云生成方法,实现从运动模糊图像中精确重建3D场景,提升实时渲染效率与重建精度。

Comments Revised Version of CRiM-GS, Project Page: https://Jho-Yonsei.github.io/CoMoGaussian

Journal ref Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2025, pp. 26415-26424

URL PDF HTML 收藏
2412.06244 2025-12-25 cs.CV

Unbiased Region-Language Alignment for Open-Vocabulary Dense Prediction

无偏区域-语言对齐用于开放词汇密集预测

Yunheng Li, Yuxuan Li, Quansheng Zeng, Wenhai Wang, Qibin Hou, Ming-Ming Cheng

机构 * VCIP, CS, Nankai University(南开大学计算机科学与技术学院) NKIARI, Shenzhen Futian(深圳福田国家信息研究院) OpenGVLab, Shanghai AI Laboratory(上海人工智能实验室) The Chinese University of Hong Kong(香港中文大学)

AI总结 DenseVLM通过无偏区域-语言对齐提升开放词汇密集预测性能

Comments Accepted at ICCV 2025. The code is available at https://github.com/HVision-NKU/DenseVLM

URL PDF HTML 收藏
2507.13387 2025-12-24 cs.CV eess.IV

From Binary to Semantic: Utilizing Large-Scale Binary Occupancy Data for 3D Semantic Occupancy Prediction

从二元到语义:利用大规模二元占用数据进行3D语义占用预测

Chihiro Noguchi, Takaki Yamamoto

机构 * InfoTech, Toyota Motor Corporation(丰田汽车公司信息科技部)

AI总结 本文提出一种基于二元占用数据的框架,通过预训练和自动标注方法提升3D语义占用预测的性能。

Comments Accepted to ICCV Workshop 2025

URL PDF HTML 收藏
2503.07940 2025-12-24 cs.CV cs.RO eess.IV

BUFFER-X: Towards Zero-Shot Point Cloud Registration in Diverse Scenes

BUFFER-X:迈向多样化场景下的零样本点云配准

Minkyun Seo, Hyungtae Lim, Kanghee Lee, Luca Carlone, Jaesik Park

机构 * Laboratory for Information & Decision Systems(信息与决策系统实验室) Massachusetts Institute of Technology(麻省理工学院)

AI总结 BUFFER-X通过自适应体素大小、最远点采样和补丁尺度归一化,实现多样化场景下的零样本点云配准,无需先验信息或手动调参。

Comments 20 pages, 14 figures. Accepted as a highlight paper at ICCV 2025

URL PDF HTML 收藏
2412.11154 2025-12-23 cs.CV

From Easy to Hard: Progressive Active Learning Framework for Infrared Small Target Detection with Single Point Supervision

从易到难:基于单点监督的红外小目标检测渐进主动学习框架

Chuang Yu, Jinmiao Zhao, Yunpeng Liu, Sicheng Zhao, Yimian Dai, Xiangyu Yue

机构 * Key Laboratory of Opto-Electronic Information Processing, Chinese Academy of Sciences(光电信息处理重点实验室,中国科学院) Shenyang Institute of Automation, Chinese Academy of Sciences(沈阳自动化研究所,中国科学院) University of Chinese Academy of Sciences(中国科学院大学) Tsinghua University(清华大学) Nankai University(南开大学) MMLab, The Chinese University of Hong Kong(香港中文大学MMLab) CPII under InnoHK(创新香港下的CPII)

AI总结 本文提出渐进主动学习框架,通过模型预启动和双更新策略提升单点监督下红外小目标检测的性能和稳定性。

Comments Accepted by ICCV 2025

URL PDF HTML 收藏
2507.23070 2025-12-23 cs.CV

Vocabulary-free Fine-grained Visual Recognition via Enriched Contextually Grounded Vision-Language Model

无需词汇的细粒度视觉识别 via 增强的上下文感知视觉语言模型

Dmitry Demidov, Zaigham Zaheer, Omkar Thawakar, Salman Khan, Fahad Shahbaz Khan

机构 * Mohamed bin Zayed University of Artificial Intelligence(莫扎德·本·扎耶德人工智能大学)

AI总结 本文提出E-FineR方法,无需词汇实现细粒度视觉识别,提供高可解释性和在零样本和少样本分类中的高性能。

Comments Accepted to ICCV 2025

URL PDF HTML 收藏
2406.05491 2025-12-23 cs.CV cs.CR

One Perturbation is Enough: On Generating Universal Adversarial Perturbations against Vision-Language Pre-training Models

一次扰动就足够:针对视觉-语言预训练模型生成通用对抗扰动

Hao Fang, Jiawei Kong, Wenbo Yu, Bin Chen, Jiawei Li, Hao Wu, Shutao Xia, Ke Xu

机构 * Tsinghua Shenzhen International Graduate School, Tsinghua University(清华大学深圳国际研究生院,清华大学) Harbin Institute of Technology(哈尔滨工业大学)

AI总结 本文提出C-PGC方法,通过对比学习生成通用对抗扰动,攻击视觉-语言预训练模型的多模态对齐能力。

Comments Accepted by ICCV-2025

URL PDF HTML 收藏