Ensembling of Deep Learning Models for Automatic Diagnosis of Breast Cancer with Histopathology Images
Gunjan Sharma, Vatsala Anand, Sheifali Gupta · 2023
Cancer-related deadly illnesses affect both developed and underdeveloped nations globally. For instance, the prevalence of breast cancer among women rises daily, partly due to ignorance and early-stage cancers not being recognized. So its early detection can help in decreasing this fatal disease. Employing computer aided diagnostic systems can help detect breast cancer at the initial level. In this research, a Convolutional Neural Network (CNN) has been proposed for the automatic detection of Breast Cancer. Firstly, ResNet50v2 pretrained model has been used for the classification task. The model has shown satisfactory results with a training accuracy of 80.17% and a validation accuracy of 81.70%. Further, the model is simulated again with the implication of the proposed CNN model. The performance of this model shows great results with a training accuracy of 82.57% and a validation accuracy of 87.80%. The value of Roc-AUC is 0.82 for this CNN model which shows the efficacy of the CNN model in classifying the images into Benign & Malignant. This research can be further used in the Medical and healthcare fields.