用支持向量机检测乳腺X线影像中的结构扭曲

The Detection of Architectural Distortion in Mammograms by Using Support Vector Machine

利用频域小波变换对乳腺X线影像中结构扭曲的图像进行小波分解,从得到的分解图像中计算了12个与纹理有关的特征参数.用支持向量机分类算法对样本集2组实验对象(乳腺结构扭曲、正常样本各19个)进行乳腺结构扭曲的识别分类;通过优化支持向量机参数条件,得到最好的分类结果.分类结果表明:本研究确定的12个纹理特征参数组合,用优化的支持向量机分类器检测和识别乳腺结构扭曲,分类正确率为92.1%、灵敏度89.5%和特异度94.7%.

The discrete wavelet transform （DWT） was employed to describe the texture features of breast in mammograms. Twelve selected parameters of the DWT were calculated and analyzed to represent the architectural distortion. Based on a sampling dataset composed of 19 architectural distortions and 19 normal mammograms, a support vector machine （SVM） was employed as a classifier. The SVM with parameters was optimized and the best classification result was obtained. The results show that the combination of all these twelve parameters by SVM is effective to detect breast architectural distortion with an accuracy of 92.1%, a sensitivity of 89.5%, a specificity of 94.7%.