Vessel Tracking Algorithms in Ultrasound Imaging
Doctors usually observe patients ultrasound vessel images to obtain clinical information about vascular diseases, and sometimes they want to measure the vessel diameter from the images. The aim of this paper is to provide a convenient method to measure the vessel diameter from the images. However, the shapes of vessels are usually anomalous, and the diameters are usually different on different part of the vessel, so it is difficult to decide the direction of the diameter as well as the length of the diameter at any given point on vessel. In this paper we proposed a Euclidean Distance Transform (EDT) based skeletonization method to find the midline of the vessel. Along the midline, we can calculate the vessel diameter locally from the intersection of vessel wall and the line perpendicular to the tangent of the midline pixels. Because of the fuzziness of the detected midline, this tangent vector is derived by a least square line fitting for robustness. Algorithms have been tested using 3D power mode kidney imaging. Results show that the proposed method can enhance the tree-like vascular structure of the kidney and offer an interactive way to measure the vessel diameter on the image.
vessel diameter Euclidean distance skeleton mathematical morphology curve fitting ultrasound image
Shuo Li Dong C. Liu
School of Computer Science Sichuan University Chengdu, Sichuan, P R China
国际会议
上海
英文
2503-2506
2008-05-16（万方平台首次上网日期，不代表论文的发表时间）