Deep Learning Implementation in the Classification of Breast Medical Images

Hassan, Elham Ahmed; Behery, Gamal; Farouk, Roshdy

doi:10.21608/bfszu.2023.199335.1262

Deep Learning Implementation in the Classification of Breast Medical Images

Document Type : Original Article

Authors

¹ Mathematics Department ,Faculty of Science,Zagazig University, Sharqie, Egypt

² Computer Information Faculty of Computer Information System Damietta University, Egypt

³ Department of Mathematics, Faculty of science, Zagazig university, Egypt

10.21608/bfszu.2023.199335.1262

Abstract

Breast cancer is one of the prime purposes of ending women's life. For this purpose, mammogram analysis is an active manner that helps radiologists in the detection of breast cancer early. This paper uses deep learning models to classify mammographic images. The support vector machine (SVM) with deep learning features of a mammogram helps to classify breast tissue based on image processing techniques. Based on the values of these features of a digital mammogram, both deep learning models and SVM try to classify the breast tissue into basic categories normal, and abnormal given in the database (mini-MIAS database). Data augmentation mechanisms have been applied to increase the training set size to avoid overfitting. After making a comparison of some models, it became clear that the best result of the classification is 97.77 % by using the VGG model. These results will be useful in making medical classification images more accurate. By this method, a radiologist can detect if the breast has cancer or not.

Keywords