With the fast increasing number of 3D models, an effective and efficient 3D model retrieval algorithm becomes more and more important. In this work, we propose a new way for extracting local features of a 3D mesh model by using tensor voting theory. Based on the new local feature descriptor, a novel algorithm for 3D model retrieval is also proposed. Firstly, a tensor voting matrix based on the normals is constructed for each vertex on the 3D mesh model. Secondly, the eigenvalues? distributions of the tensor voting matrices are used to extracting local features for the 3D model and the Bag-of-Features technique is applied to construct the feature vectors. Finally, the similarity of two 3D models is measured by the Kullback-Leibler distance. The algorithm is simple and easy to implement. Experimental results show that the algorithm is efficient and can achieve better performance when comparing with existing algorithms.
展开▼
