The CERTH-SOR3D dataset consists of over 20k instances of human-object interactions. The capturing process involved three synchronized Microsoft Kinect v2 sensors (three capturing views) under controlled environmental conditions.

The available data contain:

1. RGB videos (1920×1080 pixels ),
2. depth map sequences (512×424 pixels), and
3. 3D optical flow fields.

More details about the dataset: http://sor3d.vcl.iti.gr/