Relational Action Bank with Semantic-Visual Attention for Few-Shot Action Recognition.
Haoming LiangJinze DuHongchen ZhangBing HanYan MaPublished in: Future Internet (2023)
Keyphrases
- visual attention
- action recognition
- human actions
- recognizing human actions
- action classification
- recognizing actions
- saliency map
- higher level
- view invariant
- action primitives
- eye movements
- action detection
- recognition of human actions
- eye tracking
- video shots
- video sequences
- vision system
- activity recognition
- human detection
- bag of words
- computer vision
- visual features
- visual attention model
- space time interest points
- low level
- body parts
- atomic actions
- low level features
- human pose
- video content
- motion features
- spatio temporal
- human activities
- motion history images
- object based visual attention
- semantic similarity
- machine learning
- semantic information
- high level
- biologically inspired
- visual information
- video data
- multiscale