CLASSIFICATION OF MAMMOGRAM IMAGES

Abstract

Breast cancer is the most dangerous and incurable disease for women. Detecting breast cancer in early stages may increase the women life span. But it is not easy to identify the microcalcification in mammogram images because of its tiny size. In this paper, an author attempted to classify the mammogram images into cancer and non-cancer images by extracting low-level features. The extracted features are classified by a neural network and the SVM classifier. The neural network and the SVM classifier are trained by 150 images and tested by 50 images. The neural network accuracy is based on the mean square error and the regression analysis of the network. If the mean square error and the regression of the neural network attain low-value means, the accuracy of the network is high. Similarly, precision and recall are responsible for the SVM classifier. The database used here is the MIAS database.