Our deep learning model performs best on higher resolution images that prominently feature human subjects and natural scenery.