An LVQ-based technique for human motion segmentation.
Mochamad HariadiAkio HaradaTakafumi AokiTatsuo HiguchiPublished in: APCCAS (2) (2002)
Keyphrases
- motion segmentation
- image sequences
- motion model
- motion estimation
- missing data
- model selection
- multibody
- moving objects
- epipolar geometry estimation
- optical flow
- motion field
- spectral clustering
- flow field
- dynamic scenes
- feature tracking
- motion sequences
- multiple moving objects
- learning vector quantization
- motion compensation
- motion detection
- computer vision
- supervised learning
- three dimensional