Understanding Object Dynamics for Interactive Image-to-Video Synthesis.
Andreas BlattmannTimo MilbichMichael DorkenwaldBjörn OmmerPublished in: CVPR (2021)
Keyphrases
- image regions
- input image
- single image
- multiscale
- object motion
- spatial relationships
- region of interest
- image representation
- spatial relations
- pre trained
- complex scenes
- image segmentation
- image content
- image retrieval
- keypoints
- weakly labeled
- temporal continuity
- lighting conditions
- image data
- image features
- image frames
- normalized correlation
- successive frames
- visual data
- similar objects
- static images
- bounding box
- pixel level
- target object
- d objects
- video scene
- combining information from multiple
- multiple objects
- image set
- edge detection
- feature points
- segmentation method
- high resolution
- partial occlusion
- video surveillance
- video streams
- camera movement
- video frames
- spatial information