SiamMAST: Siamese motion-aware spatio-temporal network for video action recognition.
Xuemin LuWei QuanReformat MarekHaiquan ZhaoJim X. ChenPublished in: Vis. Comput. (2024)
Keyphrases
- action recognition
- human actions
- recognition of human actions
- recognizing human actions
- action recognition in videos
- spatio temporal
- motion features
- spatial temporal
- space time
- view invariant
- spatio temporal interest points
- static images
- human motion
- human movements
- action classification
- motion capture data
- motion history images
- human detection
- motion recognition
- action detection
- spatial and temporal
- activity recognition
- human activities
- video sequences
- bag of words
- mid level
- video dataset
- image sequences
- action primitives
- computer vision
- space time interest points
- recognizing actions
- video shots
- moving objects
- motion patterns
- human activity recognition
- motion estimation
- motion analysis
- multi view
- multimedia
- motion trajectories
- optical flow
- image classification
- visual features
- camera motion
- visual cues
- bag of features
- human pose
- temporal structure