摘要
针对冠脉造影图像噪音大且血管边缘模糊问题,提出一种基于经验模态分解的冠脉造影图像血管提取方法以较完整有效地提取出造影图像中的血管信息。该方法首先对图像预处理后的造影图像进行经验模态分解,将其分解成若干图像层,保留包含血管信息的图像层并丢弃含非血管信息图像层;其次对保留的图像层进行血管特征提取,依据连通区域的一系列属性来提取图像层中血管信息以获取清晰明显的血管段;最后通过层间信息相互参照,重构血管脉络明显的心血管图像。实验证明了方法有很高的精确度和实际的医用价值。
In this paper, we present a new vascular characteristics extraction method based on empirical mode decomposition. The proposed method overcomes the problems of noise and the blurred vascular edge generally encoun- tered in the image processing procedure. This method uses empirical mode decomposition to resolve the coronary an- giogram into several image layers, retaining the layers that contain blood vessel information and discarding layers that contain only background information. Clear and distinct vessel segments are then extracted from the reserved image layers, based on the connected region＇s properties. Finally, the coronary angiogram is reconstructed accord- ing to the information extracted from the reserved layers. The experimental results prove that this method has high accuracy and practical medical value.
出处
《智能系统学报》
CSCD
北大核心
2015年第6期851-857,共7页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金资助项目(61175053
61272369)
中央高校基本科研业务费基金主项目(2011QN 126)
关键词
血管提取
经验模态分解
冠脉造影图像
特征提取
连通区域
层间参照
血管重构
图像层
vessels extraction
empirical mode composition
coronary angiogram
characteristics extraction
connected region
interlayer reference
vascular remodeling
image layer
