Bridging the past, present and future: Modeling scene activities from event relationships and global rules.
Jagannadan VaradarajanRémi EmonetJean-Marc OdobezPublished in: CVPR (2012)
Keyphrases
- temporal relationships
- d scene
- long term
- dynamic scenes
- video sequences
- association rules
- activity recognition
- condition action rules
- scene understanding
- multiple images
- classification rules
- historical information
- global motion
- scene analysis
- complex scenes
- real world
- range images
- association rule mining
- single image
- event detection
- news articles
- human activities
- input image
- information technology
- three dimensional
- computer vision