I used modern deep-learning techniques to estimate pose and depth of subjects in videos.