Comparative Experimental İnvestigation and Application of Five Classic Pre-Trained Deep Convolutional Neural Networks via Transfer Learning for Diagnosis of Breast Cancer

In this study, for the diagnosis and classification of breast cancer, the authors used and applied five classical pre-trained deep convolutional neural network models (DCNN) which have proven successful many times in different fields (ResNet-18, AlexNet, GoogleNet and SuffleNet). In order to make pre-trained DCNN models suitable for the purpose of this study, some layers were updated according to the new situation by using the transfer learning technique. The weights of all layers used in these five pre-trained DCNN models were not changed. Instead, new weights were given to the new layers so that new layers adapt faster to the emerging new DCNN models. With these five pre-trained DCNN models, a quadruple classification as “cancer”, “normal”, “actionable” and “benign”, and a binary classification as “actionable + cancer” and “normal + benign” was realized. With these two separate classification and diagnosis studies, comparative experimental examination and analysis of pre-trained DCNN models for breast cancer diagnosis were carried out. In the study, it was concluded that successful results can be achieved with pre-trained DCNN models without extra time-consuming procedures such as feature extraction, as well as DCNN can perform quite successfully in cancer diagnosis and image comment.