Topology matching for fully automatic similarity estimation of 3D shapes
Proceedings of the 28th annual conference on Computer graphics and interactive techniques
Hi-index | 0.00 |
In content-based 3D mesh retrieval, graph-based structure is one of the most important shape descriptors. The early work on this issue can be found in [Hilaga et al. 2001], in which a 3D mesh representation, Multiresolutional Reeb Graphs (MRGs), was proposed. Since then, many Reeb-Graph based descriptors have been designed to simplify and represent 3D meshes. However, most of those descriptors are only suitable for graph-based topology matching, which is time-consuming and unreliable. To address this issue, we propose a novel approach to representing 3D mesh by using a tree-like structure, Attributed Root Trees (ARTs). The advantages of our method are three-fold: (1) Tree matching is easier and efficient than graph matching; (2) The representation well reserves topological information of a 3D mesh, thus topology matching can be easily completed; (3) It is possible to perform multi-resolution matching on ARTs.