Abstract

Info

Title:: Three-dimensional segmentation method for coronary angiographic image

Author(s):: LIU Hao1; WANG Yu1; WANG Yining2; XU Cheng2; 1. School of Artificial Intelligence, Beijing Technology and Business University, Beijing 100048, China 2. Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Beijing 100730, China

Keywords:: Keywords: coronary artery angiography three-dimensional segmentation watershed algorithm regional growth method

Abstract:: Abstract: At present, coronary computed tomography angiography has been widely used in the diagnosis of cardiovascular diseases, and the effective segmentation of coronary arteries plays an important role in the diagnosis of coronary heart diseases. Herein a three-dimensional segmentation method for coronary angiographic images is proposed. After data preprocessing, watershed algorithm is used to obtain sub-regions, and regional growth method is used to merge the sub-regions for realizing three-dimensional segmentation of coronary arteries. The manual segmentation results obtained by radiologists are used as the standard to verify the feasibility and effectiveness of the proposed algorithm. The experimental results show that the proposed segmentation method can effectively overcome the interference from the tissues around the coronary arteries and obtain accurate segmentation results.