Skip to Main Content
The 3D model matching algorithms based on the global shape features could not give the sufficient description of the local features of the model, which makes it difficult to compare models elaborately on a local level of detail. In this paper we present a new framework which combines the global topology features and the local shape features. Firstly, we employ the Potential Field Method to acquire the skeleton of the model, and then decompose it into several sub-parts according to the skeleton nodes. Secondly, we extract the local shape feature of each sub-part decomposed using the spherical harmonical method. In the process of model matching a two-step strategy has been adopted. In the first step, we compare the global topology features of the two skeletons roughly and establish the corresponding relationship of the sub-parts between the two models at the same time. In the second step, we compare the local shape features of each pair of corresponding sub-parts exactly further. Then the overall similarity of the two models consists of the weighted sum of the similarity of the global topology features and the local shape features. Experiment shows that the precision of this method is higher than the traditional algorithms based on the single shape feature or the single topology feature. Furthermore, this method can be applied to the matching based on the local feature.