Rethinking Pixel-Wise Loss for Face Super-Resolution.
Jonghyun KimJiun LeeHyeongyu KimWonjoon SongJoongkyu KimPublished in: ICEIC (2021)
Keyphrases
- super resolution
- pixel wise
- low resolution
- high resolution
- image reconstruction
- face images
- high quality
- spatial information
- pixel level
- motion estimation
- single image
- human faces
- appearance model
- motion blur
- depth map
- foreground objects
- facial expressions
- input image
- background subtraction
- template matching
- facial images
- object recognition