from the homepage: For every one hundred frames in the video, we only transmit one, and use adversarial learning to fill in the gaps. This allows us to save dat

from the homepage: For every one hundred frames in the video, we only transmit one, and use adversarial learning to fill in the gaps. This allows us to save data transmission, and makes up for latency.