Automatic Breast Cancer Detection Using Digital Thermal Images

Breast cancer is one of the most common and main reason of death for women all over the world. About one in eight of women is subject to breast cancer over the course of her lifetime. There is no effective method to prevent or know the reasons of growing these cancerous cells, however the number of deaths can be reduced by early detection. Breast cancer detection and classification is one of the most important fields that the researchers are working on. Thermal breast images are considered as an efficient type of screening strategies. The aim of this study is to develop an efficient system to detect breast cancer by using image processing techniques. The proposed system extracts the characteristic features of the breast from the region of interest that is segmented using a novel approach from the thermal input image. Then the image is classified based on these features to normal or abnormal using a neural network classifier. The system is evaluated on a benchmark dataset and a success rate of 96.51% is obtained.