View-Invariant Human Action Recognition Via View Transformation Network (VTN).
Lingling GaoYanli JiKumie GedamuXiaofeng ZhuXing XuHeng Tao ShenPublished in: IEEE Trans. Multim. (2022)
Keyphrases
- action recognition
- view invariant
- human actions
- static images
- human activities
- multiple viewpoints
- multi view
- single view
- activity recognition
- motion capture data
- body parts
- multiple views
- human detection
- computer vision
- motion history images
- human object interactions
- viewpoint
- action classification
- recognizing human actions
- image sequences
- recognition of human actions
- wireless sensor networks
- spatio temporal
- human pose
- human motion
- semi supervised
- d objects
- active learning
- feature vectors
- feature extraction