Professional Documents
Culture Documents
Make Pixels Dance- High-Dynamic Video Generation
Make Pixels Dance- High-Dynamic Video Generation
ByteDance Research
∗
Equal Contribution
arXiv:2311.10982v1 [cs.CV] 18 Nov 2023
Figure 1. Generation results of PixelDance given text, first frame instruction highlighted in red box (and last frame instruction in green).
Six frames sampled from a 16-frame clip are displayed. Human faces presented in this paper are synthesized using text-to-image models.