Deep learning techniques, one of these machine learning techniques, are also at a very important point. Thanks to the classification made with deep learning techniques, high accuracy rates can be obtained for cancer diagnosis and faster results can be obtained. In this study, VGG19 network architecture, one of the deep learning methods, was used to classify mammogram images.
In addition, image equalization and image filtering methods were applied to the images used. In this way, it was also determined which method achieved higher accuracy when the image filtering and image synchronization methods used were run together with the VGG19 deep learning network architecture. The combination of CLAHE histogram equalization and VGG19 deep learning network gave the highest accuracy. The accuracy rate of the training data in the created network is 99.82%. In addition, the loss rate of the training data in the network is 0.76% and the validation rate of the test data in the network is 99.63%. The number of correct positive images is 796, the number of false positive images is 0, the number of correct negative images is 798, and the number of false negative images is 6. These image numbers belong to the test data. The number of correctly classified images in the test data is 1594. These values are very good values for classification of mammogram images.