基于多分辨率分析的肺部边缘检测  

LUNG EDGE DETECTION BASED ON MULTI-RESOLUTION ANALYSIS

作　　者：万寿红[1] 龚育昌[1] 李曦[1] 

机构地区：[1]中国科学技术大学计算机科学技术系,安徽省计算与通讯软件重点实验室,安徽合肥230027

出　　处：《计算机应用与软件》2007年第10期109-111,共3页Computer Applications and Software

摘　　要：针对医学数字X光(DX)胸片,提出了一种基于多分辨率分析的肺部边缘检测方法。该方法结合了数学形态学和小波变换技术,构造图像的塔式分解,从低分辨率边缘图像开始,逐步向高分辨率前传、搜索、精提,自动定位肺部边缘。实验结果表明,算法计算速度快,检测出的边缘准确、连续光滑。A method of lung images edge detection is provided based on multi-resolution analysis for medical chest digital X-Ray images. The method combines mathematical morphology with wavelet transform technology. An image pyramid is constructed. Low-resolution edge image is propagated,searched and refined to high-resolution in the pyramid to detect and locate the lung edge. The experiment shows that calculation speed of the algorithm is very fast and the edge detected is accurate, smooth and continuous.

关 键 词：多分辨率分析 小波变换 数学形态学 边缘检测 DX图像 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]