融合骨架结构与几何细节的三维模型检索方法

A 3D MODEL RETRIEVAL APPROACH INTEGRATING SKELETON STRUCTURE AND GEOMETRIC DETAILS

针对拓扑方法在三维模型检索方面的不足,增加体现视觉突起和分支的关节特征点,从而有效地优化三维模型的拓扑结构。通过构建模型的拓扑结构树,并融合局部突起的几何特征,提出一种基于图结构和几何细节的模型相似性匹配算法。该方法能够弥补拓扑检索方法的单一性,实验结果验证了算法的鲁棒性与精确性。

Aiming at the weakness of topological method in 3D model retrieval, we add the articulation feature points which reflect the vis-ual protuberance and branches, so that effectively optimise the topological structure of 3D models. By constructing topological structure tree of the model and combining with the geometric characteristics of local protuberance, we further propose a similarity matching algorithm which is based on graph structure and geometric details. This method can make up the singularity of topology-based retrieval method. A series of ex-perimental results have verified the robustness and accuracy of the algorithm.