TY - GEN
T1 - Three-Dimensional Coronary Artery Centerline Extraction and Cross Sectional Lumen Quantification from CT Angiography Images
AU - Cui, Hengfei
AU - Xia, Yong
AU - Zhang, Yanning
N1 - Publisher Copyright:
© 2019, Springer Nature Switzerland AG.
PY - 2019
Y1 - 2019
N2 - Automatic centerline extraction based on 3D coronary artery segmentation results is a very important step before quantitative evaluation of intravascular lumen cross-section. In this paper, a method based on the combination of fast marching and gradient vector flow (GVF) is proposed to extract the centerline of the complete coronary artery tree in 3D angiographic images. With the centerline of blood vessel, we propose an automatic method to extract the cross-section of blood vessel lumen. This method calculates the tangent vector based on the two adjacent centerline points before and after the midline point, and then calculates the cross-sectional equation through the centerline point, and then obtains the cross-sectional contour of the cross-section and the surface mesh of blood vessel. The new method is designed to extract the cross-section of 3D intravascular lumen in real physical coordinates, which avoids the traditional interpolation processing in pixel coordinates and improves the accuracy of cross-section extraction. Given the accuracy and efficiency, the proposed coronary artery lumen area measurement algorithm can facilitate quantitative assessment of the anatomic severity of coronary stenosis.
AB - Automatic centerline extraction based on 3D coronary artery segmentation results is a very important step before quantitative evaluation of intravascular lumen cross-section. In this paper, a method based on the combination of fast marching and gradient vector flow (GVF) is proposed to extract the centerline of the complete coronary artery tree in 3D angiographic images. With the centerline of blood vessel, we propose an automatic method to extract the cross-section of blood vessel lumen. This method calculates the tangent vector based on the two adjacent centerline points before and after the midline point, and then calculates the cross-sectional equation through the centerline point, and then obtains the cross-sectional contour of the cross-section and the surface mesh of blood vessel. The new method is designed to extract the cross-section of 3D intravascular lumen in real physical coordinates, which avoids the traditional interpolation processing in pixel coordinates and improves the accuracy of cross-section extraction. Given the accuracy and efficiency, the proposed coronary artery lumen area measurement algorithm can facilitate quantitative assessment of the anatomic severity of coronary stenosis.
KW - Centerline
KW - Computer-aided diagnosis
KW - Coronary artery segmentation
KW - Cross sectional lumen area
KW - Medical imaging
UR - https://www.scopus.com/pages/publications/85077133938
U2 - 10.1007/978-3-030-36189-1_20
DO - 10.1007/978-3-030-36189-1_20
M3 - 会议稿件
AN - SCOPUS:85077133938
SN - 9783030361884
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 238
EP - 248
BT - Intelligence Science and Big Data Engineering. Visual Data Engineering - 9th International Conference, IScIDE 2019, Proceedings, Part 1
A2 - Cui, Zhen
A2 - Pan, Jinshan
A2 - Zhang, Shanshan
A2 - Xiao, Liang
A2 - Yang, Jian
PB - Springer
T2 - 9th International Conference on Intelligence Science and Big Data Engineering, IScIDE 2019
Y2 - 17 October 2019 through 20 October 2019
ER -