Breast cancer is one of the most common malignancies among women in the world, which may lead to death. Survivability rate of breast cancer can be improved if it is identified in its early stage. Breast thermography plays an important role in early detection of breast cancer, since it couples the physiological information with the anatomical features of woman breast. Typically, thermograms are visually analyzed by physicians for breast cancer early diagnosis. But it is very challenging, since it is hard to provide objective and quantitative analysis. Therefore, Computer Aided Detection (CAD) systems are used to improve the diagnostic accuracy by providing a comprehensive analysis on these Thermograms. One of the important factors that impact CAD system's performance is the classifier used for the classification of breast thermograms. However, problems such as low rate of accuracy and poor self-adaptability still exist in traditional classifiers. In this paper, a hybrid approach consists of support vector machine (SVM) classifier, with feature selection and boosting ensemble method, is proposed to enhance the performance of SVM classifier. AdaBoost algorithm is used as our boosting ensemble method. The experimental results show that the proposed hybrid approach achieves a better performance compared to the base classifier SVM alone and the base SVM classifier coupled with feature selection method. An accuracy of 99.24% is obtained using our hybrid approach. |
