2016-11-182015@inproceedings{Mejia2015_MshSegm, title = {Spectral-based Vertex Re-labeling for Mesh Segmentation}, author = {Mejia, Daniel and Ruiz, Oscar E. and Cadavid, Carlos}, booktitle = {Proceedings of the Virtual Concept International Workshop (VC-IW 2015) on Industrie 4.0}, isbn = {978-2-9548927-1-9}, month = {Nov}, year = {2015}, }978-2-9548927-1-9http://hdl.handle.net/10784/9708Mesh segmentation can be achieved by considering (a) geometric, (b) topologic or (c) a combination of geometric and topologic features on the surface Altough considering geometric characteristics would be relatively easy, our main intention is to keep the discussion on the topological aspect given that topology-based methods are foggy in their basic understanding impairing consistent application -- This article re-labels the vertices of the mesh based on the Fiedler vector (Laplacian 2nd eigenvector) for encoding the connectivity among part feature sub-meshes -- Second differences of such vector with respect to the re-labeling labeling are computed after a filter has been applied to determine the mesh partition -- The segmentation achieved by the proposed algorithm locates properly several topological features, provided the homogeneity of the triangular meshapplication/pdfenginfo:eu-repo/semantics/conferencePaperinfo:eu-repo/semantics/closedAccessANÁLISIS ESPECTRALTOPOLOGÍATRIANGULACIÓNSpectrum analysisTopologyTriangulationGeodesyMallas triangularesParámetro de FiedlerTriangulación de DelaunayAcceso cerrado2016-11-18Mejía, DanielRuíz, Óscar E.Cadavid, Carlos