光学学报, 2019, 39 (7): 0715001, 网络出版: 2019-07-16
基于角点检测的可降解支架轮廓分割算法 
下载： 823次
Corner Detection-Based Segmentation Algorithm of Bioresorbable Vascular Scaffold Strut Contours
机器视觉 角点检测 轮廓自动分割 贴壁情况分析 可降解支架 血管内光学相干断层扫描图像 machine vision corner detection contour automatic segmentation malapposition analysis bioresorbable vascular scaffold intravascular optical coherence tomography image
摘要
针对血管内光学相干断层扫描(IVOCT)影像中,使用动态规划(DP)算法进行可降解支架轮廓分割时,分割结果容易受到血液伪影和支架内部断裂的影响,而导致支架轮廓分割准确度不高的问题,利用IVOCT影像中可降解支架具有四边形外观的先验信息,提出一种使用支架的4个角点得到支架轮廓的算法。实验结果显示:所提出的支架轮廓分割算法的平均Dice系数可达到0.88,相较于DP算法提高了0.08;所提出的支架自动分割算法能够实现IVOCT影像中可降解支架的准确分割,且具有较好的稳健性,能更好地在临床应用中辅助医生进行支架贴壁情况分析。
Abstract
According to the prior knowledge about obvious quadrilateral feature of bioresorbable vascular scaffold (BVS) struts in an intravascular optical coherence tomography (IVOCT) image, this study proposes a novel algorithm based on four corners of BVS struts to automatically obtain their contours in the IVOCT imaging system. It solves the problem that dynamic programming (DP) algorithm, which is a contour-based algorithm, is not sufficiently accurate because of the in uence of the fractures inside the struts and blood artifacts around the struts. Experimental results show that the proposed algorithm achieves an average Dice's coefficient of 0.88 for the strut segmentation areas, which is increased by approximately 0.08 compared to the result obtained by the DP algorithm. This algorithm can accurately and robustly segment BVS struts in the IVOCT image, and thus it can better assist doctors in the automatic strut malapposition analysis in clinical applications.
姚林林, 金琴花, 荆晶, 陈韵岱, 曹一挥, 李嘉男, 朱锐. 基于角点检测的可降解支架轮廓分割算法[J]. 光学学报, 2019, 39(7): 0715001. Linlin Yao, Qinhua Jin, Jing Jing, Yundai Chen, Yihui Cao, Jianan Li, Rui Zhu. Corner Detection-Based Segmentation Algorithm of Bioresorbable Vascular Scaffold Strut Contours[J]. Acta Optica Sinica, 2019, 39(7): 0715001.
PDF全文
