Unsupervised Spectral Mesh Segmentation Driven by Heterogeneous Graphs.
Panagiotis TheologouIoannis PratikakisTheoharis TheoharisPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2017)
Keyphrases
- normalized cut
- image segmentation
- graph partitioning
- data driven
- segmentation algorithm
- fully unsupervised
- segmentation of natural images
- bayesian image segmentation
- multiscale
- text segmentation
- level set
- volumetric data
- segmentation method
- spectral decomposition
- segmentation of textured images
- object segmentation
- weighted graph
- spectral methods
- graph clustering
- directed graph
- graph matching
- volumetric images
- three dimensional
- graph structure
- d mesh
- edge detection
- unsupervised learning
- spectral analysis
- semi supervised
- shape prior
- region growing
- fully automatic
- medical images
- adjacency matrix
- graph mining
- hyperspectral
- energy function
- clustering algorithm