This dataset was recorded from r1-mk2.
The train folder contains 40 folders, each containing 10 repetitions of a single action. Each action is repeated 5 times and viewed from 2 points of view.
Points of view are named as following:
- v0: frontal view
- v1: side view
Actions are named as following:
- abduction-left
- external-left
- internal-left
- random (for random movements)
Each folder is named as person name-point of view - action name, containing the following subfolders:
- depth: stores depth images
- skeletons-data: stores skeletons bottles
- skeletons-img: stores skeletons images
- opc: stores skeletons from opc