MADS 人体动作数据集

MADS 数据集全称为 Martial Arts, Dancing and Sports Dataset，由香港城市大学提供。该数据集包含 5 种分类，分别为太极、空手道、爵士舞、嘻哈舞以及运动，共计 5 万 3 千帧。该数据集的动作捕捉是由两名武术大师，两名舞者和一名运动员在使用多个摄像机或立体摄像机进行录制的，可被用于人体姿态估计等领域的研究

MADS Dataset, whose full name is Martial Arts, Dancing and Sports Dataset, is provided by City University of Hong Kong. This dataset includes 5 categories, namely Tai Chi, Karate, Jazz Dance, Hip-Hop Dance and Sports, with a total of 53,000 frames. The motion capture was conducted by two martial arts masters, two dancers and one athlete, with recording performed using multiple cameras or stereoscopic cameras. It can be employed for research in fields such as human pose estimation.

搜集汇总

背景与挑战

背景概述

MADS 人体动作数据集是一个专注于武术、舞蹈和运动动作的视觉数据集，由香港城市大学发布，包含太极、空手道、爵士舞、嘻哈舞和运动5个类别，总计5.3万帧图像。该数据集通过多名专业表演者使用多摄像机系统捕捉，适用于人体姿态估计等机器视觉研究，数据规模为23.45 GB，发布于3年前。