In developing and developed nations, the risk of getting breast cancer is among the highest for women. An early diagnosis of breast cancer allows patients to be treated before metastasis (i. to I diagnoses and classi?es), detects and classi?es. This article proposed three conventional methods based on a different dataset from the rest in the literature. The models we use depend on the deep learning framework for detecting and classi?cation of breast cancer in breast X-Ray images using the concept of transfer learning. The proposed model can detect a mass region and classify it as malignant or benign on X-Ray images in one go with high accuracy. The proposed model is tested on X-Ray images from the open-source Cancer Imaging Archive (CIA) center. The data is preprocessed before giving it to the model. Transfer Learning (TL) is exploited to achieve higher prediction accuracy. Extensive simulations are conducted to measure the performance of the proposed model. The model shows that for ResNet-164 architecture, the best training and validation accuracy is 0.97% and F1 score 0.98%, and the loss rate is 0.02 at an epoch of 40. Conversely, the lowest training and validation accuracy is obtained from AlexNet with 0.91%, and F1 score 0.84%. The loss rate is 0.06 at epoch 40 for the Inception v3 and VGG19 the accuracy is 0.93%, 0.96%, and Fl score 0.89%, and 0.94%, the loss rate is 0.08, and 0.03 at epoch 40 blow-by-blow.

Username
Password