The Online Action Detection Dataset (OAD) was captured using the Kinect V2 sensor, which collects color images, depth images and human skeleton joints synchronously. This dataset includes 59 long sequences and 10 actions.
2 PAPERS • 1 BENCHMARK