Vessel intimal extraction of coronary optical coherence tomography imagery based on an improved CV model

• Main Authors: Yang, Jianli, Shi, Yasong, Lin, Feng, Yuan, Hao, Han, Yechen, Liu, Xiuling
• Other Authors: School of Computer Science and Engineering
• Format: Article
• Language: English
• Published: 2019
• Subjects: Engineering::Computer science and engineering; Active Contour; Catheter Shadow
• Online Access: https://hdl.handle.net/10356/79530; http://hdl.handle.net/10220/49061
• Institution: Nanyang Technological University

Optical coherence tomography (OCT) is a novel technology used in coronary artery disease diagnosis for its high resolution. It has numerous advantages in 3D vessel modeling and vulnerable plaque quantification. This paper proposes a robust vessel intimal extraction algorithm using active contours. A pre-processing module, that takes into consideration the characteristics of coronary OCT imagery, is introduced to eliminate a variety of visual disruptions. Then, a monotonically decreasing function is added into the Chan-Vese (CV) method in order to avoid over segmentation. This is done by controlling the evolution speed in closing the boundary area. A postprocess is followed after CV segmentation to weaken the rotation distortion influence on the extraction result. We use 281 clinical coronary OCT images gathered from Peking Union Medical College Hospital to validate our algorithm. Then we compare our results to manual expert evaluations and a conventional CV method. The extraction results can be used in clinical applications and can provide a basis to further plaque classification.