Motion saliency outweighs other low-level features while watching videos.
Dwarikanath MahapatraStefan WinklerShih-Cheng YenPublished in: Human Vision and Electronic Imaging (2008)
Keyphrases
- low level features
- key frames
- saliency map
- salient regions
- video database
- mid level
- mid level features
- low level
- visual features
- video event
- semantic gap
- video summarization
- video retrieval
- video content
- high level features
- video sequences
- low level descriptors
- video data
- high level
- video clips
- semantic concepts
- salient objects
- video indexing
- higher level
- visual cues
- high level semantics
- video frames
- visual content
- relevance feedback
- semantic attributes
- image retrieval
- feature vectors
- visual saliency
- semantic information
- image representation
- video shots
- image sequences
- visual attention
- image annotation
- low level visual features
- semantic content
- object recognition
- human motion