Are Actor and Action Semantics Retained in Video Supervoxel Segmentation?
Chenliang XuRichard F. DoellStephen Jose HansonCatherine HansonJason J. CorsoPublished in: ICSC (2013)
Keyphrases
- video segmentation
- video sequences
- human actions
- video data
- level set
- image segmentation
- video content
- segmentation method
- segmentation algorithm
- real time
- multimedia
- region growing
- object segmentation and tracking
- formal semantics
- semantic information
- multiscale
- foreground background segmentation
- action classification
- action recognition
- video surveillance
- video objects
- temporal segmentation
- text detection
- action descriptions
- video frames
- image analysis
- video database
- computer vision
- action language
- action detection
- segmentation accuracy
- video analysis
- fully automatic
- shape prior
- background subtraction
- logic programming
- object segmentation
- temporal information
- image regions
- objects in video sequences
- recognizing human actions
- spatial and temporal
- story segmentation
- atomic actions
- human activities
- key frames
- action theories
- temporal coherence
- markov random field
- motion segmentation
- video retrieval
- video clips