Trajectory-Pooled 3D Convolutional Descriptors for Action Recognition.
Xiusheng LuHongxun YaoXiaoshuai SunShengping ZhangYanhao ZhangPublished in: PCM (1) (2017)
Keyphrases
- action recognition
- action recognition in videos
- bag of features
- mid level features
- human actions
- bag of words
- activity recognition
- human detection
- computer vision
- low level descriptors
- spatial temporal
- action classification
- recognizing human actions
- body parts
- depth sensors
- feature vectors
- static images
- recognizing actions
- shape descriptors
- video dataset
- recognition of human actions
- multiscale
- action detection
- action primitives
- human activities
- three dimensional
- histogram of oriented gradients
- keypoints