Guess What Moves: Unsupervised Video and Image Segmentation by Anticipating Motion.
Subhabrata ChoudhuryLaurynas KarazijaIro LainaAndrea VedaldiChristian RupprechtPublished in: CoRR (2022)
Keyphrases
- image segmentation
- space time
- temporal filtering
- key frames
- spatial and temporal
- video sequences
- object motion
- motion features
- moving camera
- input video
- video data
- low frame rate
- motion estimation
- temporal consistency
- shot change detection
- unsupervised manner
- successive frames
- static images
- visual cues
- temporal continuity
- dynamic scenes
- motion analysis
- motion trajectories
- multimedia
- video footage
- visual data
- motion model
- video objects
- video streams
- temporal coherence
- video analysis
- surveillance videos
- normalized cut
- dynamic textures
- motion segmentation
- human motion
- single frame
- video scene
- camera motion
- unsupervised learning
- camera movement
- optical flow
- background segmentation
- computer vision
- image processing
- motion detection
- video retrieval
- image sequences
- multiscale
- level set
- video content
- deformable models
- segmentation method
- wide area motion imagery
- motion capture
- moving objects
- expectation maximization
- graph cuts
- motion capture data
- segmentation algorithm
- video frames
- video signals
- shape prior
- spectral clustering
- motion field
- markov random field
- spatio temporal
- global motion
- human actions
- motion patterns