Sapiens: Foundation for Human Vision Models
Paper
•
2408.12569
•
Published
•
89
Collect latest human image/video generation papers
Note A large model to support pose detection, depth, normal, segmentation from input human videos
Note ControlNeXt is like an improved and simplified version of classic ControlNet.
Note Preserve face ID very well in face image generation with text description control. Support multi persons. Already published demo but no code yet (08/24)