arXiv 2505.185822025-05-24cs.CVcs.AI

去噪行走视频用于步态识别

On Denoising Walking Videos for Gait Recognition

Department of Computer Science and Engineering, Southern University of Science and Technology, China（南方科技大学计算机科学与工程系）；National Engineering Laboratory for Big Data System Computing Technology, Shenzhen University, China（大数据系统计算技术国家工程实验室，深圳大学）；Shenzhen Polytechnic University, China（深圳职业技术大学）；Zhejiang University, China（浙江大学）；Alibaba Group, China（阿里巴巴集团）

Dongyang Jin, Chao Fan, Jingzhe Ma, Jingkai Zhou, Weihua Chen, Shiqi Yu

备注：: 8pages, 4 figures

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出DenoisingGait，利用生成扩散模型和几何驱动特征匹配去噪行走视频，生成步态特征场，在多个数据集上取得最先进步态识别性能。

摘要：

为了捕捉个体步态模式，在行走视频中排除与身份无关的线索（如服装纹理和颜色），仍然是基于视觉的步态识别面临的一个持久挑战。传统的基于轮廓和基于姿态的方法虽然在理论上能有效去除这类干扰，但由于其输入稀疏且信息量不足，往往难以达到高精度。新兴的端到端方法通过利用人体先验直接对RGB视频进行去噪来解决这一问题。顺应这一趋势，我们提出了DenoisingGait，一种新颖的步态去噪方法。受“我无法创造的东西，我便不理解”这一理念的启发，我们转向生成式扩散模型，揭示它们如何部分过滤掉与步态理解无关的因素。此外，我们引入了一个几何驱动的特征匹配模块，该模块结合通过人体轮廓进行的背景去除，将每个前景像素处的多通道扩散特征压缩为两通道方向向量。具体而言，所提出的帧内和帧间匹配分别捕捉步态外观和运动的局部向量化结构，产生一种新颖的类似流场的步态表示，称为步态特征场，这进一步减少了扩散特征中的残余噪声。在CCPG、CASIA-B*和SUSTech1K数据集上的实验表明，DenoisingGait在大多数情况下，在域内和跨域评估中均取得了新的最先进性能。代码可在https://github.com/ShiqiYu/OpenGait获取。

英文摘要：

To capture individual gait patterns, excluding identity-irrelevant cues in walking videos, such as clothing texture and color, remains a persistent challenge for vision-based gait recognition. Traditional silhouette- and pose-based methods, though theoretically effective at removing such distractions, often fall short of high accuracy due to their sparse and less informative inputs. Emerging end-to-end methods address this by directly denoising RGB videos using human priors. Building on this trend, we propose DenoisingGait, a novel gait denoising method. Inspired by the philosophy that "what I cannot create, I do not understand", we turn to generative diffusion models, uncovering how they partially filter out irrelevant factors for gait understanding. Additionally, we introduce a geometry-driven Feature Matching module, which, combined with background removal via human silhouettes, condenses the multi-channel diffusion features at each foreground pixel into a two-channel direction vector. Specifically, the proposed within- and cross-frame matching respectively capture the local vectorized structures of gait appearance and motion, producing a novel flow-like gait representation termed Gait Feature Field, which further reduces residual noise in diffusion features. Experiments on the CCPG, CASIA-B*, and SUSTech1K datasets demonstrate that DenoisingGait achieves a new SoTA performance in most cases for both within- and cross-domain evaluations. Code is available at https://github.com/ShiqiYu/OpenGait.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。