Two-stream RNN/CNN for action recognition in 3D videos.
Rui ZhaoHaider AliPatrick van der SmagtPublished in: IROS (2017)
Keyphrases
- action recognition
- human actions
- video dataset
- action classification
- recognition of human actions
- ucf sports
- recognizing human actions
- motion features
- recurrent neural networks
- spatio temporal interest points
- space time interest points
- human activities
- view invariant
- action detection
- static images
- activity recognition
- action recognition in videos
- recognizing actions
- bag of words
- mid level features
- human detection
- computer vision
- body parts
- data streams
- human object interactions
- spatial temporal
- motion history images
- human activity recognition
- motion capture data
- video search
- low level descriptors
- video sequences
- atomic actions
- action primitives
- video content
- human computer interaction
- mid level
- bag of features