AIST++ is a large-scale multimodal 3D dance motion dataset built from the AIST dance video database. For multi-view videos, the authors designed an elaborate pipeline to estimate camera parameters, 3D human keypoints, and 3D human dance motion sequences:

• The dataset provides 3D human keypoint annotations and camera parameters for 10.1 million images, covering 30 different subjects in 9 views. These properties make it the largest and richest existing dataset with 3D human keypoint annotations.

• The dataset contains 5.2 hours of 3D dance movements divided into 1,408 sequences covering 10 dance types, each of which includes multi-view videos shot from known camera positions and directions.