RAP 行人属性识别数据集

RAP 全称 Richly Annotated Pedestrian，是一个用于行人属性识别的数据集。它包含了 41,585 张从室内监控摄像头收集的图像，每张图像都有 72 个属性的标注，只有 51 个阳性率高于 1% 的二元属性被选作评估。该数据集训练集有 33,268 张图像，测试集有 8,317 张图像。

RAP, short for Richly Annotated Pedestrian, is a dataset dedicated to pedestrian attribute recognition. It contains 41,585 images collected from indoor surveillance cameras, with each image annotated with 72 attributes. Only 51 binary attributes with a positive rate exceeding 1% are selected for evaluation. The dataset is divided into a training set with 33,268 images and a test set with 8,317 images.

搜集汇总

背景与挑战

背景概述

RAP行人属性识别数据集包含41,585张室内监控图像，每张图像标注有72个属性，其中51个属性阳性率超过1%。数据集分为33,268张训练图像和8,317张测试图像，适用于非商业用途的行人属性识别研究。