A video processing system receives primary video capturing movement of a first object, the first object defined in the primary video by multiple contours of the first object. The video processing system receives selection information indicating a contour of interest associated with the first object; the contour of interest may be one of the multiple contours of the first object. Based on the selection information, the video processing system monitors the contour of interest. The video processing system analyzes variations in the monitored contour of interest over multiple frames of the primary video to produce magnified motion associated with the monitored contour of interest of the first object.