Facial tasks in machine learning operate based on images or video frames (or other datasets) focussed on human faces.