analyzing every frame of a video is very computationally expensive and the video resolution is often much worse than the photo resolution on such phones. But in

analyzing every frame of a video is very computationally expensive and the video resolution is often much worse than the photo resolution on such phones. But in principle the same techniques could be used so i don't see why it shouldn't be possible at some point.