Motion estimation computational complexity reduction with CNN shape segmentation.
Lauri KoskinenAri PaasioKari HalonenPublished in: IEEE Trans. Circuits Syst. Video Technol. (2005)
Keyphrases
- motion estimation
- computational complexity
- shape prior
- motion segmentation
- motion vectors
- image sequences
- image segmentation
- level set
- region based segmentation
- motion estimates
- perceptual grouping
- robust segmentation
- motion field
- motion compensation
- prior information
- energy functional
- motion compensated
- shape representation
- shape model
- medical images
- rate distortion
- segmentation method
- video coding
- video sequences
- optical flow estimation
- motion model
- geometric shapes
- cellular neural networks
- segmentation algorithm
- skeleton extraction
- optical flow
- shape variations
- shape analysis
- object shape
- video objects
- statistical shape model
- multiscale
- segmented images
- object segmentation
- shape descriptors
- deformable models
- active contours
- graph cuts
- shape decomposition
- shape constraints
- shape features
- maximum likelihood criterion
- block matching
- multiple scales
- global shape
- brain structures
- coding efficiency
- object boundaries
- energy function
- edge detection