Mesh segmentation based on face-face similarity probability

Many methods have been developed for mesh segmentation with different advantages and disadvantages. How to integrate the virtues of different segmentations is an attractive problem. This paper proposes a Face-Face Similarity Probability (FFSP) matrix for fusing the results from different segmentation algorithms. Each element of the matrix represents the probability that two faces belong to the same part. So it is a highly-condensed information package extracted from all the other algorithms´ understanding of the mesh model. After getting the FFSP matrix, we can combine it with the Random Walks (RW) algorithm for final segmentation. Compared with classical algorithms, our proposed FFSP-RW method is more competitive in both quantitative evaluation and the visualized results.