Classification of normal and abnormal patterns in medical thermography for the screening of breast cancer

Breast cancer is one of the most important causes of death among women. Breast cancer can be treated effectively only if it is detected at the early stages. Thermography plays a vital role in early detection of breast cancer. The principal aim of early detection of breast cancer is to identify the disease at a more curable stage and thus improve the prognosis and other vital clinical outcomes. In this work, we propose a classification scheme to classify the breast tissues as normal, benign or malignant and thermograms were taken from women of various age groups and various breast diseases. The proposed system consists of three stages. In the first stage pre-processing and segmentation of region of interest is performed. In the second stage, feature extraction matrix is generated using GLCM and all the detailed coefficients from 2D-DWT of a thermogram. The final stage is to classify the breast tissues with the help of Support Vector Machine (SVM), K-Nearest Neighbor (KNN). The proposed approach is applied to set of 35 images of various patients (normal, benign, malignant) and classification accuracy of 95.71%, sensitivity of 100%, specificity of 88.5%, Positive Prediction Value (PPV) of 93.62% and Negative Prediction Value (NPV) of 100% is achieved.