A novel motion-based representation for video mining.
Dong-Jun LanYu-Fei MaHong-Jiang ZhangPublished in: ICME (2003)
Keyphrases
- video representation
- space time
- layered representation
- recognizing human actions
- video streams
- temporal domain
- recognition of human actions
- spatial and temporal
- video sequences
- input video
- multimedia
- temporal filtering
- configuration space
- motion analysis
- motion capture data
- dynamic scenes
- surveillance videos
- video data
- temporal consistency
- video frames
- object motion
- key frames
- spatio temporal
- moving camera
- motion estimation
- image sequences
- motion patterns
- action recognition
- video clips
- motion capture
- visual cues
- video retrieval
- knowledge discovery
- temporal continuity
- moving objects
- data mining
- temporal structure
- optical flow
- low frame rate
- temporal coherence
- single frame
- motion features
- motion trajectories
- visual data
- video analysis
- video content
- motion segmentation
- camera motion
- motion model
- activity recognition
- multi view
- camera movement
- human motion
- video footage
- human actions
- temporal events
- successive frames
- video database
- static images
- data mining techniques