humanaigc/animateanyone

AnimateAnyone is a research project from Alibaba's HumanAIGC group that accompanies a paper titled Animate Anyone: Consistent and Controllable Image-to-Video Synthesis for Character Animation. The repository's README is essentially a landing page for that paper. The subject is image-to-video synthesis: starting from a single still image of a character and generating a video that animates them in a consistent and controllable way. The README itself does not contain code, installation instructions, or usage examples. It lists the authors of the paper, Li Hu, Xin Gao, Peng Zhang, Ke Sun, Bang Zhang, and Liefeng Bo, with links to their academic profiles. There are three badges at the top: a link to the project page on humanaigc.github.io, a link to the paper PDF on arXiv with the identifier 2311.17117, and a link to a YouTube video. There is one announcement in the README. It points readers at a follow-up open-source project from the same group called Wan-Animate, which is described as part of the Animate Anyone series. The rest of the page is a teaser image and a BibTeX citation block for academic use. Anyone looking for the actual model weights, training code, or inference scripts will not find them in this README.