Tube ConvNets: Better exploiting motion for action recognition.
Zhihao LiWenmin WangNannan LiJinzhuo WangPublished in: ICIP (2016)
Keyphrases
- action recognition
- recognizing human actions
- recognition of human actions
- feed forward
- static images
- human movements
- motion features
- action recognition in videos
- human actions
- view invariant
- action primitives
- mid level
- activity recognition
- bag of words
- computer vision
- motion capture data
- motion history images
- human detection
- action classification
- body parts
- spatial temporal
- image sequences
- human motion
- camera motion
- human activities
- optical flow
- human pose
- bag of features
- motion capture
- human activity recognition
- motion model
- recognizing actions
- motion estimation
- depth cameras
- background subtraction
- depth sensors
- space time
- view invariant action recognition