Classification of Breast Cancer Histology Images Using Transfer Learning

Abstract

Breast Cancer is a most common form of cancer among women and life taking disease around the globe. Histopathological imaging is one of the methods for cancer diagnosis where Pathologists examine tissue cells under different microscopic standards but disagree on the final decision. This is a tiresome task and for that reason, Deep Neural Networks are being used for the supervised classification. We have used Breast Histology dataset having 240 training and 20 test images for classification of the histology images among four classes, i.e. Normal, Benign, In-situ carcinoma and Invasive carcinoma. The dataset was preprocessed for proper classification. We have applied transfer learning based on AlexNet, GoogleNet, and ResNet that can classify images at multiple cellular and nuclei configurations. This approach has resulted in 85% accuracy in case of ResNet as the highest among others and further research is being done to increase its efficiency and reduce the human dependency. The proposed design can also be enhanced for automation of other medical imaging methods.