ContourRender: Detecting Arbitrary Contour Shape For Instance Segmentation In One Pass.
Tutian TangWenqiang XuRuolin YeYanfeng WangCewu LuPublished in: CoRR (2021)
Keyphrases
- shape prior
- shape constraints
- geometric active contours
- dice similarity coefficient
- active contours
- global shape
- contour detection
- deformable contour
- shape description
- image segmentation
- geometric shapes
- boundary extraction
- level set
- closed contours
- external force
- object contours
- ellipse fitting
- object regions
- automatic initialization
- energy functional
- intensity information
- segmentation method
- shape descriptors
- object boundaries
- image gradient
- shape model
- geometrical features
- image contours
- contour points
- deformable models
- shape variations
- segmented images
- closed curves
- multiscale segmentation
- segmentation algorithm
- force field
- deformable template
- fourier descriptors
- shape features
- graph cuts
- region homogeneity
- region boundaries
- multiscale
- mr images
- contour segments
- contour matching
- level set evolution
- gradient vector flow
- polygonal approximation
- chain code
- shape analysis
- statistical shape model
- object segmentation
- active contour model
- evolution equation
- seed points
- maximum likelihood criterion
- region growing
- shape representation
- vector field
- arbitrary shape
- perceptual grouping
- edge strength
- object shape
- medical images
- edge detection
- level set method
- contour fragments
- energy function
- contour extraction