Enhancing Breast Cancer Detection Through a Tailored Convolutional Neural Network Deep Learning Approach

Round the globe, the common form of malignancy dreadful disease settling in women is breast cancer. The disparate impact of breast cancer in women makes it an enthralling topic for research studies. Besides, the prediction of breast cancer is imperative and essential to conceive a clear schema of therapy and personalized medication. Moreover, breast cancer has a profound emotional and psychological impact on patients and their support networks, emphasizing the need for comprehensive care and support. As a result of the disease’s costs of diagnosis, treatment, and lost productivity, healthcare systems and economies are burdened due to the high costs of therapy. This paper contributes to developing a deep learning model, Custom CNN to categorically classify breast cancer images into benign and malignant forms by analyzing the histopathological images of breakHis dataset. Over and above, the results obtained are further compared with the additional pre-trained models MobileNetV3, EfficientNETB1, VGG16, and ResNet50V2 with the same dataset. Among them, Custom CNN stimulated an accuracy’s ultimate common form of malignancy of 92% which outperforms the other CNN models.