3 ms·
i guess an interesting question is: can you coax a network into learning a better perceptual compression transform than the dcts in jpeg?
by a-dub 3y ago
i guess an interesting question is: can you coax a network into learning a better perceptual compression transform than the dcts in jpeg?