Decouple Content and Motion for Conditional Image-to-Video Generation.
Cuifeng ShenYulu GanChen ChenXiongwei ZhuLele ChengTingting GaoJinzhi WangPublished in: AAAI (2024)
Keyphrases
- static images
- image frames
- object motion
- visual data
- temporal continuity
- visual cues
- successive frames
- key frames
- image data
- image features
- space time
- single frame
- video footage
- multimedia
- normalized correlation
- image segmentation
- multiscale
- image retrieval
- input video
- input image
- feature points
- visual concepts
- single image
- video sequences
- image motion
- video scene
- motion estimation
- video streams
- video data
- image content
- camera movement
- multimedia data
- image classification
- motion analysis
- textual descriptions
- displacement estimation
- temporal domain
- web images
- layered representation
- reference frame
- high resolution
- motion segmentation
- human motion
- motion field
- spatial and temporal
- low level
- computer vision
- multi frame
- moving objects
- camera motion
- structure from motion
- motion model
- visual information
- video analysis
- motion trajectories
- video shots