Multi-view transition HMMs based view-invariant human action recognition method.
Xiaofei JiZhaojie JuCe WangChanghui WangPublished in: Multim. Tools Appl. (2016)
Keyphrases
- view invariant
- multi view
- action recognition
- single view
- human actions
- camera calibration
- computer vision
- multiple viewpoints
- pairwise
- depth map
- multiple views
- motion capture data
- semi supervised
- spatio temporal
- viewpoint
- range images
- human motion
- hidden markov models
- surface reconstruction
- geometric constraints
- face tracking
- invariant representation
- face recognition
- three dimensional