中文摘要：

针对大部分骨架计算方法对轮廓噪声的极端敏感性问题,该文提出一种基于贝叶斯模型的骨架裁剪方法。该方法利用贝叶斯理论对骨架及其生长过程进行建模,进而通过对模型的迭代优化实现骨架候选分支的筛选裁剪。由于已有的重建误差率在分析骨架时不能很好地体现骨架简洁程度,故该文在骨架重建误差率的基础上综合考虑骨架简洁度,提出骨架有效率的概念来对骨架做客观定量分析。实验结果表明该文算法对轮廓噪声具有较好的鲁棒性,且裁剪出的骨架相比现有算法得到的骨架结构更加简单,对形状描述更加准确。

英文摘要：

Considering the problem that most of the existing skeleton calculation methods exhibit extreme sensitivity to the shape noise, a Bayes based algorithm for the skeleton pruning is proposed. The algorithm models the skeleton and growth process with Bayesian statistics framework. Based on the model, an iterative optimization is performed to prune the candidate branches. Due to the fact that the existing reconstruction error can not evaluate the simplicity of skeletons well, a new concept called Effective Rate is proposed to make quantitative analysis on the pruned skeleton with taking the simplicity into consideration. The experiments show that the proposed algorithm is robust to the shape noise and acts better in simplifying the skeleton structure and representing shape accurately.