Action Recognition in Untrimmed Videos with Composite Self-attention Two-Stream Framework.
Dong CaoLisha XuHaibo ChenPublished in: ACPR (2) (2019)
Keyphrases
- action recognition
- video dataset
- human actions
- recognition of human actions
- action classification
- activity recognition
- bag of words
- static images
- spatial temporal
- recognizing human actions
- human activities
- human detection
- view invariant
- motion features
- recognizing actions
- computer vision
- space time interest points
- action detection
- spatio temporal
- human pose
- body parts
- video clips
- event detection
- video frames
- motion history images
- text classification
- data streams
- spatio temporal interest points
- video sequences