基于区域增长法的三维支气管树分割算法

Three Dimensional Airway Trees Segmentation Algorithm Based on Region Growing Method

虚拟内窥镜在胸部疾病诊断方面占据重要的地位,通常支气管树分割方法存在分割结果不准确和分割漏洞问题,因此提出基于区域增长法的支气管树分割算法。首先通过区域增长法进行主干分割,然后对细小分支进行获取,并通过质量评价函数对细小分支进行筛选,删除伪分支。实验结果表明,该支气管树分割法可以简单、有效地提取出完整的肺支气管树,得到包含第5级以上的支气管,解决支气管断裂和分割漏洞现象,具有较好的鲁棒性。

Virtual bronchoscopy（VB） plays an important role in the evaluation of chest diseases, but as the main key technology, airway tree segmentation method has problems of inaccurate and leakage phenomenon. Therefore, the airway trees segmentation based on region growing method was proposed. Firstly, region growing method is used to extract the main branch. Secondly, it extracts the sub-branch and deletes the pseudo branch according to the quality evaluation func tion. The experiment results show that the robust airway segmentation method can extract the complete lung airway tree simply,robustly and effectively with magnitude 5 bronchus,which solves the bronchial rupture and segmentation leakage problems.