By collecting these blood flow signals from the entire face and converting them into a spatiotemporal map by the algorithm, deep learning can immediately detect whether the image is real or fake.