Video dataset of 102 participants for the paper "Learning deep representations for video-based intake gesture detection"