CMU Panoptic (CMU Panoptic Dataset)

CMU Panoptic是一个大规模数据集，为多个人参与社交活动提供3D姿势注释 (1.5百万)。它包含65个视频 (5.5小时) 与多视图注释，但其中只有17个是在多人场景和相机参数。
大规模多视图系统
480 VGA相机视图
30个高清视图
10个RGB-D传感器
基于硬件的同步
校准
带有标签的有趣场景
多人
社会互动群体
3D身体姿势
3D面部地标
成绩单演讲者ID
硬件设置
480 VGA相机，640x480分辨率，25 fps，使用硬件时钟相互同步
31个高清摄像机，1920x1080分辨率，30 fps，使用硬件时钟相互同步，与VGA摄像机对齐的时间
10个Kinect Ⅱ 传感器。1920x1080 (RGB)，512x424 (深度)，30 fps，它们与其他传感器之间的定时对齐5 DLP投影仪。与高清摄像机同步

CMU Panoptic is a large-scale dataset that provides 1.5 million 3D pose annotations for social activities involving multiple people. It contains 65 videos (totaling 5.5 hours) with multi-view annotations, but only 17 of them feature multi-person scenes and calibrated camera parameters.

Large-scale multi-view system:
- 480 VGA camera views
- 30 HD camera views
- 10 RGB-D sensors
- Hardware-based synchronization
- Calibration

Labeled engaging scenarios:
- Multi-person scenarios
- Social interaction groups
- 3D body poses
- 3D facial landmarks
- Transcribed speaker IDs

Hardware setup:
1. 480 VGA cameras: 640×480 resolution, 25 fps, synchronized with each other via hardware clocks
2. 31 HD cameras: 1920×1080 resolution, 30 fps, synchronized via hardware clocks and time-aligned with the VGA cameras
3. 10 Kinect Ⅱ sensors: 1920×1080 (RGB) resolution, 512×424 (depth) resolution, 30 fps, time-aligned with other sensors; 5 DLP projectors, synchronized with the HD cameras