StyleHEAT: One-Shot High-Resolution Editable Talking Face Generation via Pre-trained StyleGAN.
Fei YinYong ZhangXiaodong CunMingdeng CaoYanbo FanXuan WangQingyan BaiBaoyuan WuJue WangYujiu YangPublished in: ECCV (17) (2022)
Keyphrases
- pre trained
- high resolution
- face images
- low resolution
- training data
- super resolution
- remote sensing
- training examples
- high quality
- human faces
- control signals
- three dimensional
- pairwise
- image processing
- principal component analysis
- active learning
- prior knowledge
- feature vectors
- face recognition
- appearance variations
- image sequences