Human action recognition based on quaternion spatial-temporal convolutional neural network and LSTM in RGB videos.
Bo MengXue-jun LiuXiaolin WangPublished in: Multim. Tools Appl. (2018)
Keyphrases
- spatial temporal
- action recognition
- human actions
- convolutional neural network
- human activities
- static images
- human object interactions
- motion history images
- action classification
- motion capture data
- video dataset
- recognition of human actions
- space time interest points
- bag of words
- recognizing human actions
- activity recognition
- computer vision
- video database
- view invariant
- recognizing actions
- action recognition in videos
- action detection
- face detection
- human pose
- video shots
- bag of features
- video frames
- space time
- depth data
- video data
- video sequences
- image sequences
- machine learning