Encoding Multi-resolution Two-Stream CNNs for Action Recognition.
Weichen XueHaohua ZhaoLiqing ZhangPublished in: ICONIP (3) (2016)
Keyphrases
- action recognition
- multiresolution
- bag of words
- human actions
- data streams
- activity recognition
- computer vision
- human detection
- spatial temporal
- body parts
- action classification
- independent subspace analysis
- recognizing human actions
- view invariant
- mid level
- static images
- video dataset
- recognizing actions
- stream processing
- human activities
- depth sensors
- video sequences
- human pose
- bag of features
- motion features
- viewpoint
- action primitives
- multiscale