Accurate 3D face reconstruction via prior constrained structure from motion.
Matthias HernandezTal HassnerJongmoo ChoiGérard G. MedioniPublished in: Comput. Graph. (2017)
Keyphrases
- structure from motion
- bundle adjustment
- image sequences
- computer vision
- missing data
- point cloud
- feature tracking
- camera parameters
- camera motion
- scene structure
- high quality
- depth information
- reconstruction from multiple images
- relative pose
- reprojection error
- feature correspondences
- camera positions
- focal length
- human faces
- camera pose estimation
- image based localization
- camera pose
- line correspondences
- key ingredient
- d objects