ByteDance's OmniHuman-1 model is able to create realistic videos of humans talking and moving naturally from a single still image, according to a paper published by researchers with the tech company.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback