A Pre-Trained Ensemble Model for Breast Cancer Grade Detection Based on Small Datasets

Background and Purpose: Nowadays, breast cancer is reported as one of the most common cancers amongst women. Early detection of the cancer type is essential to aid in informing subsequent treatments. The newest proposed breast cancer detectors are based on deep learning. Most of these works focus on large-datasets and are not developed for small datasets. Although the large datasets might lead to more reliable results, collecting and processing them are challenging. Materials and Methods: This paper proposes a new ensemble deep learning model for breast cancer grade detection based on small datasets. Our model uses some basic deep learning classifiers for grading the breast tumors including grades I, II and III. Since none of the previous works focus on the datasets including breast cancer grades, we have used a new dataset, called Databiox, for grading the breast cancers in the three grades. Databiox includes histopathological microscopy images of Invasive Ductal Carcinoma (IDC) diagnosed patients. Results: The performance of the model is evaluated based on the small dataset. We compare the proposed three layers’ ensemble classifier with the most common single deep learning classifiers in terms of accuracy and loss. The experimental results show that the proposed model could improve the classification accuracy of the breast cancer grade in comparison to the other state-of-the-art single classifiers. Conclusion: The ensemble model can be also used for small datasets. In addition, they can improve the accuracy compared with the other models. This achievement is principal for designing classification-based systems in computer-aided diagnosis.