论文部分内容阅读
冠脉光学相干断层成像(OCT)图像斑块区域分割是冠脉斑块识别的前提和基础,对后续斑块特征分析及易损斑块识别,进而实现冠脉疾病的辅助诊断分析具有十分重要的意义。本文提出了一种新的算法,使用K-means算法与图割算法结合,实现了冠脉OCT图像斑块准确的多区域分割——纤维化斑块、钙化斑块和脂质池,并较好地保留了斑块的边界特征信息。本文实验中对20组具有典型斑块特征的冠脉OCT图像进行了分割,通过与医生手动分割结果比较,证明本文方法能准确地分割出斑块区域,且算法具有较好的稳定性。研究结果证明了本文工作能够极大减少医生分割斑块所消耗的时间,避免不同医生之间的主观差异性,或可辅助临床医生对冠心病的诊断与治疗。
Coronary Optical Coherence Tomography (OCT) Image Plaque region segmentation is the prerequisite and foundation of coronary plaque recognition. It is very important for the follow-up plaque feature analysis and vulnerable plaque identification to further diagnose coronary artery disease Meaning. In this paper, we propose a new algorithm that combines the K-means algorithm with the graph cut algorithm to achieve accurate multislice segmentation of coronary OCT images - fibrotic plaques, calcified plaques and lipid pools Good retention of the plaque boundary feature information. Twenty groups of coronary OCT images with typical plaque characteristics were segmented in this experiment. Compared with the manual segmentation results of doctors, it is proved that this method can accurately segment the plaque region and the algorithm has good stability. The results of the study demonstrate that the work in this paper can greatly reduce the time spent by doctors in dividing plaque, avoid subjective differences among different doctors, or assist clinicians in the diagnosis and treatment of coronary heart disease.