Video-Processing: Map Body to Face

[rohansaw]

In order to nicer visualize results, map the identification we performed to the according track of bounding-boxes for the gorilla body.

We are able to create continuous track of bboxes for the body but for the face only when we have a frontal view. Therefore, identify the body-bbox that a face lies in and add the identity of that face to the body track and visualize in the output video.