A skeleton pruning algorithm based on information fusion

Skeleton pruning is an essential part of the processing and analysis of skeletons. It is still quite a challenging problem because of the lack of standard measurements for the importance or significance of a branch. The relative significance of the same branches will be different if we see them from different perspectives with different objectives. Different objective measurements have their advantages and limitations. To integrate the advantages of different objective measurements, we consider skeleton pruning as a multi-objective decision-making problem and propose a skeleton pruning algorithm based on information fusion. During the pruning process, we use combinatorial fusion analysis and the concept of cognitive diversity to fuse various measurements of branch significance including region reconstruction, contour reconstruction and visual contribution. Experimental results show that: (1) the proposed method is stable across a wide range of shapes and robust to boundary noise, and (2) it can effectively generate multi-scale skeletons according with visual judgment. (C) 2013 Elsevier B.V. All rights reserved.