This is wild.
Microsoft just dropped VASA-1.
And it can make any image sing and talk, with just an audio 
12 wild examples (+link at the end)
1. Mona Lisa rapping
12 wild examples (+link at the end)
1. Mona Lisa rapping 2. Real-time generation
3. Same motion sequence with different photos
4. Example of eye gaze direction and head distance, and emotion offsets
5. Different emotion offsets
6. Pose and expression editing
7. Realism and liveliness
8. Singing
9. Realism and liveliness - 2
10. Realism and liveliness - example 3
11. Realism and liveliness - example 4
12. Realism and liveliness - example 3
Read more at the official Microsoft Research blog:
microsoft.com/en-us/research
Sai Rahul
@sairahul1
Sharing insights on Al 🤖 Bootstrapping https://t.co/zYLVAN4UM9 - fastest way to grow your Insta 🚀
Missing some tweets in this thread? Or failed to load images or videos? You can try to .