-LND: A two-stream framework with discriminant deep descriptor, linear CMDT and nonlinear KCMDT descriptors for action recognition.
Ming TongMengao ZhaoYiran ChenHouyi WangPublished in: Neurocomputing (2019)
Keyphrases
- action recognition
- bag of features
- action recognition in videos
- video dataset
- activity recognition
- bag of words
- human actions
- computer vision
- human detection
- spatial temporal
- body parts
- recognition of human actions
- action classification
- probabilistic model
- action detection
- recognizing human actions
- recognizing actions
- low level descriptors
- human computer interaction
- human activities
- human pose
- static images
- data streams