Motion-transformer: self-supervised pre-training for skeleton-based action recognition.
Yi-Bin ChengXipeng ChenDongyu ZhangLiang LinPublished in: MMAsia (2020)
Keyphrases
- action recognition
- recognizing human actions
- recognition of human actions
- static images
- human movements
- action recognition in videos
- motion features
- view invariant
- human actions
- action primitives
- activity recognition
- mid level
- bag of words
- motion capture data
- action classification
- body parts
- human detection
- human motion
- spatial temporal
- motion history images
- image sequences
- computer vision
- bag of features
- motion model
- space time
- motion estimation
- motion capture
- motion analysis
- multi view
- optical flow
- spatio temporal
- recognizing actions
- view invariant action recognition
- motion recognition
- human activity recognition
- motion patterns
- spatial and temporal
- robust tracking
- depth map
- three dimensional