基于小波变换的骨肉瘤CR图像分析  

Asteosarcoma CR Images Analysis Based on Wavelet Transform

作　　者：胡珊[1] 张俊杰[1] 徐超[1] 刘燕[1] 

机构地区：[1]中山大学中山医学院计算机中心,广州市中山二路74号510060

出　　处：《中国数字医学》2014年第6期71-74,共4页China Digital Medicine

摘　　要：目的:研究骨肉瘤影像诊断的数字化特征指标。方法:对71例正常骨和70例骨肉瘤CR图像,按病变区域分为骨端和骨干两部分,运用Db4小波和Sym4小波两种小波变换方法,提取出目标区域的纹理特征,并通过统计学方法筛选出特征子集,采用支持向量机的方法构建骨肉瘤CR图像的计算机分类模型,并对模型进行正确性、敏感性和特异性的验证分析。结果:分类模型的正确性、敏感性和特异性的验证结果表明,对骨端骨肉瘤,Sym4小波的识别正确率为93.44%,对骨干骨肉瘤,Db4小波的识别正确率为96.25%。结论:小波变换法所提取的纹理特征对识别正常骨和骨肉瘤影像有较好的意义,有助于在计算机辅助诊断中构建骨肉瘤的数字化诊断标准。Objective：To find and validate texture features extracted from bone CR images used in computer-aided osteosarcoma diagnosis systems.Methods：The bone CR images were divided into two classes,epiphysis and diaphysis.Sym4 and Db4 wavelet transforms were applied to extract texture features from these images respectively.To obtain the optimal set of features,statistical methods were used to maximize the feature selection.Then a support vector machine algorithm was used to evaluate the performance of these methods.Results：Accuracy,sensitivity and specificity were used to evaluate the recognition rate of osteosarcoma.The experimental results demonstrated that the Sym4 wavelet had a higher classification accuracy （93.44％） than the Db4 wavelet with respect to osteosarcoma occurrence in the epiphysis,whereas the Db4 wavelet had a higher classification accuracy （96.25％） for osteosarcoma occurrence in the diaphysis.Conclusion：This study confirms that multi-resolution analysis is a useful tool for texture feature extraction during bone CR image processing.A set of texture features can be extracted from the wavelets and used in computer-aided osteosarcoma diagnosis systems.

关 键 词：骨肉瘤 纹理特征 小波变换 特征选择 

分 类 号：R318.1[医药卫生—生物医学工程]