Market1203-Reid-Dataset

Market-1203 数据集：该数据集包含从两个不相交的摄像机视图中捕获的 1203 个人。对于每个人，在一个相机视图下从一到六个不同方向捕获一到十二个图像，并标准化为 128x64 像素。该数据集是基于 Market-1501 基准数据构建的，我们手动注释每个图像的方向标签。我们随机选择 601 个人进行训练，其余人进行测试。Market-1203 Dataset: This dataset consists of 1203 individuals captured from two disjoint camera views. For each individual, 1 to 12 images are captured from 1 to 6 distinct directions within a single camera view, and all images are standardized to a resolution of 128×64 pixels. This dataset is built upon the Market-1501 benchmark, and we manually annotated the orientation labels for every image. We randomly select 601 individuals as the training subset, with the remaining ones forming the test subset.

搜集汇总

背景与挑战

背景概述

Market1203-Reid-Dataset是一个用于行人重识别任务的图像数据集，包含来自两个摄像机视角的1203个行人图像，每张图像已标准化为128x64像素并标注了方向标签。该数据集基于Market-1501构建，其中601人用于训练，其余用于测试。