Search

Breast cancer is the most common cancer among women worldwide. Early-stage diagnosis of this disease can significantly improve the efficiency of treatment. Computer-Aided Diagnosis (CAD) Systems are adopted widely in this regard due to their reliability, accuracy and affordability. There are different imaging techniques for a breast cancer diagnosis; one of the most accurate ones is histopathology which is used in this paper. Deep feature transfer learning is used as the main idea of the proposed CAD system's feature extractor. As such, the present paper works on sixteen different pre-trained networks with a focus on their classification phase, something that has not been studied enough. The Inception-ResNet-v2, which has both residual and inception networks profits together, has shown the best feature extraction capability in the case of breast cancer histopathology images among all tested Convolutional neural networks (CNNs). In the classification phase, the ensemble of Categorical Boosting (CatBoost), Extreme Gradient Boosting (XGBoost) and Light Gradient boosting Machine (LightGBM) has given the best average accuracy. The Breast Cancer Histopathological Image Classification (BreakHis) dataset helps evaluating the proposed method, i.e., IRv2-CXL, with the experimental results indicating that IRv2-CXL outperforms other state-of-the-art methods. [Display omitted] • A novel deep feature transfer learning model, called IRv2-CXL, is proposed. • IRv2-CXL is implemented to improve prediction accuracy and lower variance. • IRv2-CXL is used for breast tumours classification in histopathological images. • Over 100 different structures are evaluated to identify the best model. • IRv2-CXL outperformes other state-of-the-art methods on the BreakHis dataset. [ABSTRACT FROM AUTHOR]