A Novel Coronary Artery Skeleton Extracting Method Based on Multiple Hypothesis Tracking

作者：Zhu, Wen Bo; Li, Bin^*; Tian, Lian Fang; Chen, Kan

来源：Journal of Medical Imaging and Health Informatics, 2014, 4(6): 832-839.

DOI：10.1166/jmihi.2014.1332

• Skeleton Extracting
• 3D Coronary Artery Segmentation
• Local Shape Analysis
• Multiple Hypothesis Tracking
• Cardiac Image

Skeleton extracting plays an important role in many fields, including medical image processing, bioinformatics, pattern recognition and others, and the research of inherent topological structure of object shapes is therefore needed. Most existing approaches, applied on the coronary artery images, face the problems of low accuracy, high computation cost and lack of radius information. In this paper, a novel skeleton extracting method for 3D coronary artery is proposed. The proposed method is capable of extracting skeleton robustly and automatically by employing a new medialness measuring function, level-set graph and multiple-hypothesis tracking strategy. It also involves a pre-segmentation step for the goal of reducing undetermined image noise disturbance. The proposed skeleton extracting method is consistent and scalable for vessels of different sizes, and has been validated on both synthetic and clinical images. Experimental results show that the proposed method outperforms most existing algorithms.

• 出版日期2014-12
• 单位华南理工大学