基于核图割算法的冠脉光学相干断层图像斑块区域分割

发布时间：2018-04-09 21:18

本文选题：斑块区域分割　切入点：冠脉光学相干断层成像图像　出处：《生物医学工程学杂志》2017年01期

【摘要】：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。
[Abstract]:Plaque segmentation in optical coherence tomography (Oct) image of coronary artery is the premise and foundation of coronary plaque recognition, which is of great significance to the analysis of plaque characteristics and vulnerable plaque recognition, and to the diagnosis and analysis of coronary artery disease.In this paper, a new algorithm is proposed, which combines the K-means algorithm with the graph cutting algorithm to realize the accurate multi-region segmentation of coronary artery OCT image plaques-fibrosis plaque, calcified plaque and lipid pool, and better preserve the plaque boundary feature information.In this paper, 20 groups of coronary artery OCT images with typical plaque characteristics were segmented. Compared with the results of manual segmentation by doctors, it is proved that the proposed method can segment the plaque region accurately and the algorithm has good stability.The results show that this work can greatly reduce the time spent by doctors to segment plaque, avoid subjective differences between different doctors, or assist clinicians in the diagnosis and treatment of coronary heart disease.
【作者单位】：河北大学电子信息工程学院;中国医学科学院北京协和医院心内科;
【基金】：国家自然科学基金资助项目(61473112) 河北省杰出青年科学基金资助项目(F2016201186)
【分类号】：R541.4;TP391.41

【参考文献】