search

magic-research / magic-animate

[CVPR 2024] MagicAnimate: Temporally Consistent Human Image Animation using Diffusion Model
https://showlab.github.io/magicanimate/
BSD 3-Clause "New" or "Revised" License
10.4k stars 1.07k forks source link

Choice of densepose #40

Open smandava98 opened 10 months ago

smandava98 commented 10 months ago

Hi. Thanks for the great work. What was the reasoning behind the choice for densepose? I see that the segmentation masks are not always accurate so it leads to some inconsistencies in human biomechanics. Out of curiosity, would such a diffusion model work if one were to map the body onto a body model like SMPL and go from there?

chen-rn commented 10 months ago

Curious as well. It seems that Densepose doesn't quite capture the finger/hand movements.