文章摘要

文章信息/Info

作者:: 杨泽鹏1; 3; 李娜2; 张保昌2; 吴宗翰2; 杨俊3△; 周寿军2△; 1. 华南理工大学广州学院，广州 510800；2. 深圳先进技术研究院，深圳 518055； 3. 火箭军广州特勤疗养中心，广州 510515

Author(s):: YANG Zepeng1; 3; LI Na2; ZHANG Baochang2; WU Zonghan2; YANG Jun3; ZHOU Shoujun2; 1. Guangzhou College of South China University of Technology, Guangzhou 510800,China ; 2. Shenzhen Institutes of Advanced Technology, Shenzhen 518055, China; 3. The Rockets Military Nursing Home，The PLA Rocket Force，Guangzhou 510515

Keywords:: Angiography image segmentation; Model-driven approach; Data-driven approach; Machine learning method; Deep learning method

摘要:: 血管分割能够为计算机临床辅助与可视化诊疗提供关键的结构和病灶信息。由于血管造影图像涉及多种数据模态以及个体化差异和复杂病理，诸多挑战性的问题仍待解决。我们对血管分割的主流方法进行了回顾和分析，并且讨论了血管分割的研究格局和发展趋势，旨在为该领域的进一步研究提供参考。

Abstract:: Segmentation of angiographic images can provide the key structure and focus information for visual diagnosis and treatment with computer clinical assistant. Related researches have been carried out for decades. Since angiographic images involve multiple data modalities as well as individualized differences and complex pathologies, many challenging issues remain to be resolved. We review and analyze the main methods of vascular segmentation, as well as discuss the research pattern and development trend of vascular segmentation to provide reference for the further researches in this field.