人体行为图像&视频描述数据

20,000 张人体行为图像&视频描述数据,包含20,000张图像，10,000段视频采集了不同季节，不同拍摄角度的多种人体行为，包含室内场景与室外场景。描述语言为英语，主要描述了人物性别、年龄、衣着、行为描述与肢体动作。

This dataset is a human behavior image and video caption dataset, which includes 20,000 images and 10,000 video clips. It captures diverse human behaviors across different seasons and shooting perspectives, covering both indoor and outdoor scenarios. The captioning language is English, and it mainly describes the subject's gender, age, attire, behavioral characteristics and body movements.

搜集汇总

背景与挑战

背景概述

该数据集包含20,000张图像和10,000段视频，专注于人体行为描述，采集自不同季节、拍摄角度和室内外场景，使用英语标注人物性别、年龄、衣着及行为动作。它是商业数据集，仅限企业合作购买使用。