Most existing person re-identification (re-id) models focus on matching still
person images across disjoint camera views. Since only limited information can
be exploited from still images, it is hard (if not impossible) to overcome the
occlusion, pose and camera-view change, and lighti