This is an example of one video from our ShadowPunch dataset with coresponding annotation.