Skeleton-based segmentation and decomposition of raster pairs of shapes
Liudmila Domakhina · Pattern Recognition and Image Analysis · 2010
The problem of discrete shape and its skeleton decomposition is considered in this work—so-called skeleton-based shape segmentation problem. An approach to construct the isomorphic skeleton-based decompositions for pairs of shapes is proposed. Based on these isomorphic decompositions of skeletons, isomorphic shapes decompositions with meaningful parts reflecting shape structure could be constructed. The last feature provides an opportunity to apply the proposed method to shape comparison and shape morphing problems. Experiments confirm the correctness of methods proposed.