LEO: Generative Latent Image Animator for Human Video Synthesis.
Yaohui WangXin MaXinyuan ChenAntitza DantchevaBo DaiYu QiaoPublished in: CoRR (2023)
Keyphrases
- static images
- input image
- multiscale
- image data
- image classification
- image content
- test images
- visual data
- image features
- texture synthesis
- video data
- high resolution
- image frames
- image analysis
- image segmentation
- single image
- human observers
- multimedia
- edge detection
- low level
- image retrieval
- spatial information
- image representation
- video frames
- object recognition
- video sequences
- visual attributes
- segmentation method
- image regions
- feature points
- video content
- image collections
- weakly labeled