Experimental Study of a Deep-Learning RGB-D Tracker for Virtual Remote Human Model Reconstruction

Figure 4

Tracked stick-figure information and point cloud segmentation: (a) overlaid tracked skeleton model on live visual image and (b) segmented point clouds associated with the corresponding definition of the skeleton joints and limbs. Information from such segmented point clouds is further utilized to develop a novel refinement of the tracked information associated with the locations of the skeleton joints.

(a)

(b)