Motion-modulated Temporal Fragment Alignment Network For Few-Shot Action Recognition.
Jiamin WuTianzhu ZhangZhe ZhangFeng WuYongdong ZhangPublished in: CVPR (2022)
Keyphrases
- action recognition
- recognizing human actions
- motion history images
- recognition of human actions
- static images
- human actions
- spatial temporal
- motion features
- human movements
- action recognition in videos
- space time
- activity recognition
- view invariant
- spatial and temporal
- mid level
- action primitives
- video sequences
- human detection
- human motion
- computer vision
- temporal continuity
- bag of words
- motion capture data
- action classification
- body parts
- spatio temporal
- motion estimation
- key frames
- human activity recognition
- bag of features
- human activities
- video shots
- moving objects
- depth sensors
- machine learning
- temporal structure
- visual features
- motion trajectories
- motion analysis
- temporal resolution
- multi view
- object detection
- robust tracking
- optical flow
- human pose
- image sequences
- motion patterns