Typefully

Animate-Anyone: Image-to-Video Synthesis for Character Animation

Avatar

Share

 • 

2 years ago

 • 

View on X

This is Animate-Anyone A new Image-to-Video synthesis for character/person animation by researchers from Alibaba Group Lets have a look, videos and links below 🧵 A Thread
This is really impressive, and you can just imagine how you can take a selfie do a virtual try-on and get some really nice short snippets.
You can also generate dance videos of your completely AI generated influencer that you are secretly managing.
If thats not enough, you can animate characters without really having to have 3D models of them. One image is enough.
Just look at this, standard pose for a 3D character model, and then it can be manipulated like this. 🤯 Mind blown.
Check out the full paper here humanaigc.github.io/animate-anyone/
And check out the youtube video here youtube.com/watch?v=8PCn5hLKNu4
Avatar

Linus Ekenstam

@LinusEkenstam

AI Evangelist & Optimist. Latest AI News, Trends and learn how to use AI tools to augment your abilities. eu/acc 🇪🇺 Co-founder @flocurve @bedtimestoryAI