3D skeleton based action recognition by video-domain translation-scale invariant mapping and multi-scale dilated CNN.
Bo LiMingyi HeYuchao DaiXuelian ChengYucheng ChenPublished in: Multim. Tools Appl. (2018)
Keyphrases
- action recognition
- scale invariant
- multiscale
- human actions
- action classification
- scale space
- natural images
- action detection
- recognition of human actions
- human activities
- keypoints
- recognizing human actions
- bag of words
- affine invariant
- activity recognition
- computer vision
- local binary pattern
- space time interest points
- video sequences
- image structure
- motion history images
- texture images
- human pose
- video frames
- invariant features
- video data
- spatio temporal
- image processing
- human motion
- denoising
- feature detection
- image features
- multiresolution
- image segmentation