基于拉曼光谱和SVM的乳腺病灶识别模型研究

Breast Disease Recognition Model Research Based on Raman Spectroscopy and Support Vector Machine

乳腺癌是女性主要癌症之一,若癌细胞进一步转移到骨骼、中枢神经系统和内脏,将会导致预后不良和总体生存率的降低。相比于传统的诊断乳腺肿瘤的病理学方法耗时且破费的特点,拉曼光谱的检测方法损伤较小且诊断周期短。本文利用吉林大学第一医院乳腺外科提供的实验检测样本,建立了新鲜乳腺病灶组织的拉曼光谱数据库,在特征选择的基础上应用支持向量机(SVM)方法构建了乳腺组织良恶性识别模型,并运用集成学习的思想以便快速鉴别乳腺病灶的类型。

Breast cancer is one of the leading cancers in women, if the cancer cells further transfer to the bones and internal organs, central nervous system will result in poor prognosis and the overall survival rate lower. Compared with the traditional pathological methods, Raman spectroscopy method is time-consuming and expensive. In this paper, a Raman spectral database of fresh breast lesions was established by using the experimental test samples provided by the department of breast surgery, the first hospital of Jilin University. On the basis of feature selection, a benign and malignant breast tissue recognition model was established by using support vector mechanism as well as ensemble learning in order to quickly identify the types of breast lesions.