Precision Breast Cancer Detection with Advanced Neural Network Methods

Precision Breast Cancer

Abstract

Breast cancer remains a significant health issue for women, and while
mammography is widely used, it still presents diagnostic challenges. Recent
advancements in deep learning, particularly convolutional neural networks
(CNNs), offer promising solutions for improving mammographic image
classification. This manuscript presents a method using CNNs to classify
tumors in mammographic images from the DDSM database. Comparing an
artificial neural network (ANN) and an EfficientNet CNN, both trained on
Google Colab, the ANN achieved an accuracy of 98%, while the EfficientNet
CNN achieved 94%. Additionally, the model demonstrated 99% accuracy in
classifying masses as "spiculated," indicating potential malignancy, and 99%
accuracy in classifying masses as "obscured," suggesting a possible aggressive
tumor. These results highlight the potential of deep learning models to provide
accurate diagnostics and assist radiologists in breast cancer treatment. The
study also focuses on optimizing hyperparameters to further enhance detection