论文部分内容阅读
目的利用乳腺肿瘤超声图像良恶性的不同特征,借助于模式分类方法对乳腺肿瘤良恶性进行识别,作为医生的计算机辅助诊断。方法本文研究基于乳腺肿瘤超声图像的原始特征参数已提取情况下,采用顺序前进搜索方法获得最优特征矢量,然后利用支撑矢量机、贝叶斯分类器、BP网络和Fisher线性判别器四种模式识别方法分别对乳腺肿瘤良恶性进行识别。结果基于200例病例随机划分为训练集100例和测试集100例进行测试,支撑矢量机、贝叶斯分类器、BP网络和Fisher线性判别器的Accuracy分别为0.960,0.940,0.932±0.013,0.930。结论支撑矢量机的分类性能优于其它分类器,能有效地对超声图像乳腺肿瘤进行良恶性识别。
Objective To use the different features of benign and malignant ultrasound images of breast tumors and to identify the benign and malignant breast tumors with the help of pattern classification method. Methods In this paper, based on the extraction of the original feature parameters of ultrasound images of breast tumors, a sequential forward search method was used to obtain the optimal feature vectors, and then four modes were used: support vector machine, Bayesian classifier, BP network and Fisher linear discriminator. Recognition methods identify benign and malignant breast tumors, respectively. The results were based on 200 cases randomly divided into 100 training sets and 100 test sets. The Accuracy of the support vector machine, Bayesian classifier, BP network, and Fisher linear discriminator were 0.960, 0.940, 0.932±0.013, and 0.930, respectively. . Conclusion The classification performance of the support vector machine is better than other classifiers, and can be used to identify benign and malignant ultrasound mammary tumors effectively.