基于形态学的冠状动脉造影图像多尺度增强算法

提出了一种基于形态学的多分辨率增强算法,将心脏冠状动脉数字造影图像序列的连续帧相减,区分血管所在区域与背景区域,然后利用得到的区域信息通过形态学方法对冠状动脉血管进行多分辨率增强。采用该方法处理后的心脏冠状动脉造影图像的背景被减弱,血管被增强,产生类似数字减影图像的效果。实验表明,增强后的图像质量良好,对临床诊断和血管分割具有较高实用价值。

Morphological multiscale vessel enhancement for coronary angiograms

A vessel enhancement algorithm was developed for coronary angiography based on the morphological multiscale principle.Areas containing vessels were distinguished from the background by subtraction of sequential frames.The vessel visibility relative to surrounding structures was increased by enhancing a morphological multiscale structure.The vessels were enhanced using a nonlinear function that is stronger at locations likely to contain vessels.The coronary angiograms after processing were given a DSA-like,vessel enhanced appearance.Tests show that the algorithms is indeed valuable for diagnosis and vessel extraction.