1 paper · 1 filter
Riku Takahashi, Ryugo Morita, Fuma Kimishima +2
Existing deep facial animation coding techniques efficiently compress talking head videos by applying deep generative models. Instead of compressing the entire video sequence, thes…