DICE: End-to-end Deformation Capture of Hand-Face Interactions from a Single Image.
Qingxuan WuZhiyang DouSirui XuSoshi ShimadaChen WangZhengming YuYuan LiuCheng LinZeyu CaoTaku KomuraVladislav GolyanikChristian TheobaltWenping WangLingjie LiuPublished in: CoRR (2024)
Keyphrases
- end to end
- single image
- partially occluded
- shape recovery
- multiple images
- d scene
- image pairs
- light source
- shape from shading
- face images
- super resolution
- admission control
- congestion control
- image registration
- human faces
- intrinsic images
- application layer
- computer vision
- d morphable model
- transport layer
- face recognition