BREAST CANCER DETECTION USING DEEP LEARNING ALGORITHMS

Abstract

Breast cancer is a common malignancy among women and ranks as one of the most common malignant malignancies after lung cancer. Breast cancer detection is based on mammography film. Mammography films can be used to diagnose breast cancer in the female population. However, mammograms do not allow accurate diagnosis of breast cancer, resulting in misdiagnosis. Hence, a critical need arises for an integrated system employing a deep learning approach to comprehensively classify breast cancer type, sub-type, and grade. The implementation of such a system holds the potential to alleviate the substantial workloads of pathologists and mitigate the risk of misdiagnoses. The MATLAB software provides several functions for machine learning algorithms and image processing of breast tumor images. The study demonstrates improved accuracy compared to established algorithms and models. The algorithm employs Swish, LeakyReLU, ReLU, and Sigmoid activation functions for activation. The model achieves high accuracy rates for benchmark datasets and avoids overfitting, incorporating multiple variations for CNN training algorithms.