Gaze location prediction for broadcast football video using Bayesian integration of low level features and top-down cues.
Qin ChengDimitris AgrafiotisAlin AchimDavid R. BullPublished in: ICIP (2013)
Keyphrases
- low level features
- mid level
- video database
- high level
- key frames
- location prediction
- semantic concepts
- low level
- video data
- visual cues
- news video
- higher level
- image retrieval
- visual features
- semantic attributes
- high level semantics
- semantic information
- video content
- visual information
- image representation
- relevance feedback
- video sequences
- saliency map
- video retrieval
- video shots
- video streams
- temporal information
- multimedia
- eye tracking
- object recognition
- video clips
- video frames
- hot spots
- visual saliency
- video analysis
- link prediction
- multimedia data
- space time
- contextual information
- image classification
- metadata
- multimedia databases
- machine learning
- eye movements
- domain knowledge
- query processing
- mobile objects
- computer vision
- social networks
- search engine