Human action recognition based on multi-scale feature maps from depth video sequences.
Chang LiQian HuangXing LiQianhan WuPublished in: Multim. Tools Appl. (2021)
Keyphrases
- action recognition
- human actions
- feature maps
- multiscale
- motion capture data
- video sequences
- depth sensors
- body parts
- human activities
- space time interest points
- activity recognition
- motion history images
- action classification
- motion recognition
- human motion
- human object interactions
- kernel function
- neural network
- self organizing maps
- saliency map
- computer vision
- image representation
- spatio temporal
- space time
- feature vectors
- depth map
- human visual system
- video data
- natural images
- relevance feedback
- high quality
- image sequences
- three dimensional
- recognizing human actions
- multimedia
- feature selection