LitNeRF: Intrinsic Radiance Decomposition for High-Quality View Synthesis and Relighting of Faces.
Kripasindhu SarkarMarcel C. BühlerGengyan LiDaoye WangDelio ViciniJérémy RiviereYinda ZhangSergio Orts-EscolanoPaulo F. U. GotardoThabo BeelerAbhimitra MekaPublished in: SIGGRAPH Asia (2023)
Keyphrases
- view synthesis
- depth map
- high quality
- illumination conditions
- reference images
- lighting conditions
- spectral reflectance
- human faces
- multi view
- real scenes
- face recognition
- depth information
- d scene
- image based rendering
- intermediate views
- stereo reconstruction
- high resolution
- face images
- multiple cameras
- image quality
- scene geometry
- viewpoint
- depth estimation
- three dimensional
- image sequences
- photorealistic
- ground truth
- light field
- stereo pair
- highly accurate
- d objects
- stereo matching
- super resolution
- light source
- low resolution
- dynamic scenes
- disparity map
- dynamic programming
- depth images
- camera views
- partial occlusion
- input image
- object recognition
- virtual view
- face detection