Publication

Abstract : Breast cancer is a life-threatening disease that affects individuals all over the world. As a consequence, effective and precise breast cancer screening is crucial. Early detection of breast cancer allows patients to obtain the best treatment available, boosting their probability of surviving. Several studies have resulted in the development of computational algorithms for predicting breast cancer progression using a variety of diagnostic imaging modalities. In this paper, a deep learning approach for the classification of breast cancer histopathology images is carried out. The proposed model is hybrid combination of Inspection-ResNetv2 and EfficientNetV2-S with pretrained weights as ImageNet. The proposed model was validated on BreakHis and Breast Cancer Histology (BACH) dataset. For concatenation of both networks, top layer was removed and global average pooling was added, followed by dense layer, dropout and final classification layer. The proposed model was evaluated in comparison with individual results obtained by Inspection-ResNetv2 and EfficientNetV2 model results. The final classification layer consists of dense layer of four neurons of BACH dataset classification and for BreakHis 8 neurons. The experimentation of proposed model showed good results by achieving overall accuracy of 98.15% for the BACH dataset and 99.036% for BreakHis dataset.