Self-supervised High-fidelity and Re-renderable 3D Facial Reconstruction from a Single Image.
Mingxin YangJianwei GuoZhanglin ChengXiaopeng ZhangDong-Ming YanPublished in: CoRR (2021)
Keyphrases
- single image
- high fidelity
- shape recovery
- high resolution
- super resolution
- image pairs
- multiple images
- high quality
- real time
- light source
- medical image compression
- shape from shading
- d scene
- facial expressions
- image reconstruction
- indoor scenes
- face recognition
- three dimensional
- human faces
- face images
- vanishing points
- facial images
- reconstruction method
- facial features
- intrinsic images
- outdoor images
- fundamental problem in computer vision