NTU RGB+D 2D

本页面介绍两个数据集: "NTU RGB D" 和 "NTU RGB D 120"。
“NTU RGB D” 包含60个动作类和56,880个视频样本。
“NTU RGB D 120” 通过添加另外60个类和另外57,600个视频样本来扩展 “NTU RGB D”，即 “NTU RGB D 120” 总共具有120个类和114,480个样本。
这两个数据集都包含每个样本的RGB视频，深度图序列，3D骨骼数据和红外 (IR) 视频。每个数据集由三个Kinect V2摄像机同时捕获。
RGB视频的分辨率为1920x1080，深度图和IR视频均为512x424，3D骨骼数据包含每帧25个身体关节的3D坐标。

This page introduces two datasets: "NTU RGB-D" and "NTU RGB-D 120".
"NTU RGB-D" consists of 60 action categories and 56,880 video samples.
"NTU RGB-D 120" expands the "NTU RGB-D" dataset by adding another 60 categories and 57,600 additional video samples, resulting in a total of 120 action categories and 114,480 samples for "NTU RGB-D 120".
Both datasets contain RGB videos, depth map sequences, 3D skeleton data, and infrared (IR) videos for each sample. Each dataset was captured simultaneously by three Kinect V2 cameras.
The resolution of RGB videos is 1920×1080, while both depth maps and IR videos have a resolution of 512×424. The 3D skeleton data includes 3D coordinates of 25 body joints per frame.