BCTD: An Optimised CNN-based Deep Neural Network for Primary Stage Breast Cancer Detection and Classification
Sharif Hasan, Md. Mahadi Hassan Nour, Abdullah Al Kafi, Md Fakrul Islam, Muhammad Nazrul Islam, Faiz Al Faisal · 2024
This research focuses on breast cancer detection and classification, which is a global health concern and affects millions of people all around the world.Accurate diagnosis and type classification of breast cancer is essential for creating a successful treatment strategy and guaranteeing patient recovery, while that is an open-ended complex engineering problem.This study proposes an optimized deep neural network intended for breast cancer primary-stage detection and classification by utilizing cutting-edge convolutional neural network (CNN) techniques.Our model achieves an impressive training accuracy of 98.92%, validation accuracy of 97.55%, and test accuracy of 95.76% on the BreakHis dataset, while 96% F1-score for both cancer types.Moreover, we have compared our research analysis with the existing research works, which shows an important development over the existing research on breast cancer detection and classification.