基于多尺度空间滤波和l_1范数最近邻分类的乳腺图像微钙化点检测  

The microcalcifications detection of mammograms based on multi-scale space filtering and l_1 norm nearest-neighbor classifier

作　　者：胡正平[1] 刘敏华[1] 

机构地区：[1]燕山大学信息科学与工程学院,河北秦皇岛066004

出　　处：《电路与系统学报》2011年第1期85-91,共7页Journal of Circuits and Systems

基　　金：国家自然科学基金(61071199);河北省自然科学基金(F2008000891;2010001297);中国博士后自然科学基金(20080440124);第二批中国博士后基金特别资助(200902356)

摘　　要：钙化信息是乳腺癌早期诊断的一个重要依据,针对钙化点检测检出率较低和假阳性较高的问题,提出一种基于多尺度空间滤波和l1范数最近邻分类的乳腺图像微钙化点检测算法。首先利用多尺度空间滤波方法得到原图像的多尺度显著特征图,然后通过基于人眼视觉特性的钙化点分割方法得到粗检测钙化点的二值图像,并送入l1范数最近邻分类器去除假阳性点。仿真实验结果表明,本文的钙化点粗检测算法可较好的检测出可疑钙化点,对于对比度较低的钙化图像也可得到较好的检测结果,同时后续的分类器判决效果良好,使钙化点检测结果具有较高检出率的同时具有较低的假阳性率。The microcalcification information is an important foundation for the diagnosis of breast cancer. In order to improve the problem of true-positive and false-positive in microcalcifications detection, a novel microcalcifications detection algorithm of rnammograms based on multi-scale space filtering and l1norm nearest-neighbor （l1-NN） classifier is proposed. Firstly the multi-scale salience feature images are obtained by using multi-scale space filtering for original images, then the coarse detected binary image of microcalcifications is induced via using microcalcifications segmentation method based on human visual model., and into the l1-NN to remove the false-positive points. Simulation results demonstrate that the proposed coarse detection method can effectively detect the suspicious microcalcifications from the mammograms including low contrast images, and the following classifier has a good performance. The microcalcifications detection has higher true-positive rate and lower false-positive rate.

关 键 词：微钙化点检测 多尺度变换 l1范数最近邻分类器 乳腺图像 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]