CASIA-E: A Large Comprehensive Dataset for Gait Recognition

Gait recognition plays a special role in visual surveillance due to its unique advantage, e.g., long-distance, cross-view and non-cooperative recognition. However, it has not yet been widely applied. One reason for this awkwardness is the lack of a truly big dataset captured in practical outdoor scenarios. Here, the “big” at least means: (1) huge amount of gait videos, (2) sufficient subjects, (3) rich attributes, and (4) spatial and temporal variations. Moreover, most existing large-scale gait datasets are collected indoors, which have few challenges from real scenes, such as the dynamic and complex background clutters, illumination variations, vertical view variations, etc. In this paper, we introduce a newly built big outdoor gait dataset, called CASIA-E. It contains more than one thousand people distributed over near one million videos. Each person involves 26 view angles and varied appearances caused by changes of bag carrying, dressing and walking styles. The videos are captured across five months and across three kinds of outdoor scenes. Soft biometric features are also recorded for all subjects including age, gender, height, weight and nationality.[News!] Now, our dataset is supported by the popular gait recognition project OpenGait. We highly recommend that you can adopt the new setting in (https://github.com/ShiqiYu/OpenGait/blob/master/datasets/CASIA-E/README.md) as the default setting.If you use our dataset in your research, please cite the corresponding paper:Chunfeng Song1, Yongzhen Huang1, Weining Wang, and Liang Wang*, "CASIA-E: A Large Comprehensive Dataset for Gait Recognition," in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), vol. 45, no. 3, pp. 2801-2815, 1 March 2023, doi: 10.1109/TPAMI.2022.3183288. (The first two authors contribute equally to this work)BibTex:@article{casiae2022,title={CASIA-E: a large comprehensive dataset for gait recognition},author={Song, Chunfeng and Huang, Yongzhen and Wang, Weining and Wang, Liang},journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},year={2022},volume={45},number={3},pages={2801-2815},publisher={IEEE}}

步态识别（gait recognition）因其独特优势在视觉监控领域占据特殊地位，例如可实现远距离、跨视角与非配合式识别。但目前其应用尚未得到广泛普及。造成这一窘境的原因之一，是缺乏真正在实际户外场景下采集的大型数据集。此处“大型”至少需满足四点要求：（1）海量步态视频数据；（2）充足的受试者数量；（3）丰富的属性维度；（4）多样的时空变化条件。
此外，当前多数已有的大规模步态识别数据集均在室内环境采集，难以应对真实场景中的诸多挑战，例如动态复杂背景杂波、光照变化、俯视角变化等。本文介绍了全新构建的大型户外步态识别数据集CASIA-E。该数据集涵盖千余名受试者，对应近百万条步态视频。每位受试者对应26个采集视角，且包含因携带物品、着装与行走姿态变化所产生的多样化外观差异。视频采集周期跨越五个月，覆盖三类户外场景。所有受试者的软生物特征（soft biometric features）均已记录，包括年龄、性别、身高、体重与国籍。
【新动态！】目前，本数据集已获得主流步态识别项目OpenGait的官方支持。强烈建议您采用该项目文档（https://github.com/ShiqiYu/OpenGait/blob/master/datasets/CASIA-E/README.md）中提供的默认配置。
若您在研究中使用本数据集，请引用以下论文：宋春峰1，黄永桢1，王卫宁，王亮*，"CASIA-E：面向步态识别的大型综合数据集"，载于《IEEE模式分析与机器智能汇刊》（IEEE Transactions on Pattern Analysis and Machine Intelligence，TPAMI）第45卷第3期，第2801-2815页，2023年3月1日，DOI: 10.1109/TPAMI.2022.3183288。（本文前两位作者贡献均等）
BibTex:@article{casiae2022,title={CASIA-E: a large comprehensive dataset for gait recognition},author={Song, Chunfeng and Huang, Yongzhen and Wang, Weining and Wang, Liang},journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},year={2022},volume={45},number={3},pages={2801-2815},publisher={IEEE}}

搜集汇总

背景与挑战

背景概述

CASIA-E是一个大型户外步态识别数据集，包含1000多人的近百万段视频，涵盖多种视角和外观变化，并记录了丰富的软生物特征。数据集支持OpenGait项目，适用于步态识别研究。