摘要
血管内超声(IVUS)图像冠状动脉血管壁的边缘提取对冠状动脉疾病的诊断和治疗有着重要意义。本研究提出了一种用于自动提取IVUS序列图像冠状动脉血管壁内、外膜边缘的方法。该方法基于活动轮廓模型以及本研究所定义的边缘对比度特征量,利用Hopfield神经网络并结合模拟退火算法自动提取IVUS序列图像冠状动脉血管壁的内、外膜边缘。实验结果表明,本研究方法易于实现,而且准确性和可靠性较高,对IVUS序列图像处理的可重复性和鲁棒性较好,是一种较好的全局最优化算法。
The edge detection of intravascular ultrasound （IVUS） images is an important issue in the diagnosis and treatment of the coronary artery disease. A method for detecting automatically the intima and adventitia edge of IVUS images was proposed in this paper. The method was based on the active contour model and the edge-contrast, and that Hopfield neural network was employed as well as the simulated annealing algorithm. The experimental results showed that the method was simple, accurate, reproducible and robust for sequential IVUS frames, and was a kind of global optimal algorithm.
出处
《中国生物医学工程学报》
CAS
CSCD
北大核心
2006年第4期385-389,395,共6页
Chinese Journal of Biomedical Engineering
基金
国家自然科学基金资助项目(60271015)
卫生部临床学科重点项目(20012943)
山东省优秀中青年科学家科研奖励基金资助项目(2005BS01006)
