Fig. 2From: Flexible human action recognition in depth video sequences using masked joint trajectoriesHuman body joints used in our method. Numbering and names correspond to the ones used in the skeleton trackerBack to article page