A Critical Analysis of Deep Learning Architectures for Classifying Breast Cancer Using Histopathology Images

Cancer is the classification given to a group of diseases that occurs when abnormal cells divide uncontrollably, often destroying normal, healthy tissue. Cancer which is a genetic disease is often caused by the change in the genetic makeup of living cells. Medical research has identified and classified over 100 subcategories of cancer, with the names given to each being derived from the organ where the cell mutation occurs. Medical professionals often utilize pathology reports to aid in the diagnosis and treatment of cancer; these reports contain a vast amount of information and include histopathology scans. The article uses varying techniques to learn the patterns found in histopathology scans so that the use of deep learning architectures in the classification of cancerous tissue can be compared and analysed. The results of the research have shown that the AlexNet model achieves an accuracy of 79%, the DenseNet model achieves an accuracy of 84% while the NASNet Mobile model achieves an accuracy of 88%.