行人重识别测试数据Market1501数据集

Market1501数据集主要面向重识别领域，满足行人重识别的研究需求。数据集包含了32668张图片，共 1501 个身份。这些行人图片是由 6 个监控摄像头（其中5个高清摄像头和1个低清摄像头）收集而来，每个行人至少由2个摄像头捕获到，并且在一个摄像头中可能具有多张图像。
随后分为训练子集 12,936 张图片，共 751 个身份，平均每个人有17.2张训练数据；测试子集 19,732 张图片，共 750 个身份，平均每个人有26.3张测试数据。查询集有3368张查询图像。图片分辨率统一为128X64。

The Market1501 dataset is primarily targeted at the person re-identification domain, meeting the research requirements of pedestrian re-identification studies. It contains a total of 32,668 images, corresponding to 1,501 unique pedestrian identities. These pedestrian images were collected by 6 surveillance cameras, including 5 high-definition cameras and 1 low-definition camera. Each pedestrian is captured by at least two cameras, and multiple images of the same pedestrian may exist within the footage captured by a single camera.

The dataset is subsequently split into three subsets: training, testing, and query. The training subset consists of 12,936 images, covering 751 unique identities, with an average of 17.2 training images per identity. The testing subset has 19,732 images, involving 750 unique identities, with an average of 26.3 testing images per identity. The query subset includes 3,368 query images. All images have a unified resolution of 128×64.

搜集汇总

背景与挑战

背景概述

Market1501数据集是一个用于行人重识别研究的大规模测试数据集，包含32668张图片，覆盖1501个不同身份，由6个监控摄像头采集，确保每个行人至少出现在两个摄像头中。该数据集分为训练和测试子集，图片分辨率统一为128x64，适用于计算机视觉和人工智能领域的模式识别任务。