Recognizing activities and poses: lessons from computer vision