CHIRLA

该数据集名为CHIRLA，专为长期人员重识别任务设计。它包含了来自战略位置摄像机连续七个月的记录，捕捉了时间和外观属性的重大变化。数据集涵盖了22位个体，其中包括17名男性和5名女性。视频以AVI格式存储，标注则以JSON格式提供。规模上，该数据集拥有超过5.52小时的视频时长，以及大约一百万个边界框。所涉及的任务是人员重识别。

The dataset, named CHIRLA, is specifically designed for the long-term person re-identification task. It consists of recordings captured by strategically positioned cameras over a continuous seven-month period, capturing significant variations in both temporal and appearance attributes. The dataset covers 22 individuals, including 17 males and 5 females. The videos are stored in AVI format, while the annotations are provided in JSON format. In terms of scale, this dataset has a total video duration of over 5.52 hours and approximately one million bounding boxes. The targeted task of this dataset is person re-identification.