Özet:

Breast cancer is a type of cancer that usually occurs in the lobule, duct and connective tissue regions of the breast and is caused by the abnormal movement of cells in these regions. It is among the most common types of cancer among women in general. When the disease is detected early, cancerous cells can affect other organs through the blood and lymphatic vessels (metastasis state). Therefore, early diagnosis and treatment of breast cancer is important. In this study, an artificial intelligence-based early diagnosis system that can classify between benign and malignant types of breast cancer is proposed. Residual blocky neural network models are used in the proposed approach. Type-based features were extracted by adding a new fully connected layer to the last layer of ResNet models. In the next step, a new feature set was created by combining the features obtained from the fully connected layers. Softmax and machine learning methods (support vector machines, nearest neighbor method, etc.) were used in the classification process. With the proposed approach, 100% overall accuracy was obtained from all the methods used in the classification process. In this study, it was observed that obtaining and combining type-based fully connected layers positively affected the performance of experimental analyzes.

Anahtar Kelimeler: