3D-Hog Embedding Frameworks for Single and Multi-Viewpoints Action Recognition Based on Human Silhouettes.
Federico AngeliniZeyu FuSergio A. VelastinJonathon A. ChambersSyed Mohsen NaqviPublished in: ICASSP (2018)
Keyphrases
- action recognition
- motion capture data
- view invariant
- static images
- bag of words
- human actions
- human movements
- human activities
- human detection
- activity recognition
- d objects
- action classification
- body parts
- viewpoint
- bag of features
- computer vision
- spatial temporal
- recognizing actions
- human pose
- recognizing human actions
- human object interactions
- motion history images
- visual hull
- action detection
- action primitives
- recognition of human actions
- histogram of oriented gradients
- video dataset
- multi view