Abstract: Computer vision advancements allow motion transfer for animating static objects in images. However, current methods rely on manually collected motion labels and struggle with accurate shape ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback