مجلة جامعة الزيتونة الدولية للنشر العلمي مجلة علمية، محكمة، شهرية، مفتوحة الوصول
Abstract
Breast cancer is one of the primary causes of death among women. Early detection of breast cancer allows for the receipt of appropriate treatment, thus increasing the possibility of survival. In this paper, we proposed a hybrid deep learning model using a pre-trained VGG16 model with a self-attention mechanism for breast cancer detection. We extract features from the binary class (benign, malignant) dataset of the mammographic image analysis society (MIAS) using pre-trained deep convolutional neural network (CNN) architectures like Xception, MobileNet, DenseNet, and VGG-16. So the results illustrated that the best model is VGG16 with a self-attention module, which achieved an accuracy of 98.77%.
Keywords: Breast cancer, VGG16, MIAS, Mammography, Classification.
BY :
Tawfik Ezat Mousa¹, Mohamed S. Geoda²
¹² Departement of Computer Technologies,
Higher Institute Of Science and Technology, Tobruk, Libya.
