DeepChange

DeepChange是一个大规模的长时人员再识别基准数据集，专注于衣物变化。该数据集由牛津大学和萨里大学创建，包含1,121个身份和178,407个边界框，覆盖了长达12个月的时间跨度。数据集通过17个户外分辨率不同的摄像头在真实世界的监控系统中录制，捕捉了高度多样的衣物变化和风格，以及从分钟到季节不同的时间间隔。DeepChange数据集的应用领域主要集中在解决长时人员搜索中的衣物变化问题，为研究提供了丰富的真实场景和挑战。

DeepChange is a large-scale long-term person re-identification benchmark dataset focused on clothing variation. Developed by the University of Oxford and the University of Surrey, this dataset contains 1,121 identities and 178,407 bounding boxes, spanning a time period of up to 12 months. Recorded via 17 outdoor cameras with varying resolutions in real-world surveillance systems, it captures highly diverse clothing variations and styles, as well as time intervals ranging from minutes to seasons. Its primary application scenarios focus on addressing the clothing variation problem in long-term person search, providing rich real-world scenarios and challenges for relevant research.