View-invariant action recognition using cross ratios across frames.
Yeying ZhangKaiqi HuangYongzhen HuangTieniu TanPublished in: ICIP (2009)
Keyphrases
- view invariant
- cross ratios
- action recognition
- invariant representation
- human actions
- multi view
- single view
- cross ratio
- human motion
- projective invariants
- viewpoint
- multiple viewpoints
- video frames
- geometric structure
- affine transformation
- projective transformations
- error analysis
- gesture recognition
- computer vision
- human activities
- activity recognition
- visual features
- moving objects
- multiple views
- range images
- space time
- human body
- spatio temporal