Homotopy Semilattices

[摘要] A new, quasi-self-dual approach for morphological image processing is introduced. This approach is based on a complete semilattice framework. The related morphological erosion, for instance, shrinks connected components in an image, regardless to whether they are bright or dark. In the binary case, the morphological operators are based on a complete inf-semilattice that is related to the homotopy tree. Two grayscale generalizations are investigated. The first is based on the topographic-distance concept, whereas the second one utilizes the reconstruction-based fillhole operation. 18 Pages

[发布机构] HP Development Company

[学科分类] 计算机科学（综合）

[关键词] mathematical morphology;complete inf-semilattices;erosion;opening;topographic distance;reconstruction;fillhole