Comparison of DenseNet201 and ResNet50 for Breast Cancer Detection using Deep Learning

Abstract

Breast cancer is one of the most prevalent and life- threatening diseases affecting women worldwide. Early detection plays a crucial role in improving survival rates and treatment effectiveness. Recently, deep learning has emerged as a powerful tool in medical imaging, particularly in the classification and detection of breast cancer. In this study, we compare the performance of DenseNet201 and ResNet50 architectures on a breast ultrasound image dataset for classification tasks. The models were trained for 20 epochs, achieving test accuracies of 83.17% and 65.42% for DenseNet201 and ResNet50, respectively. The corresponding training accuracies were 90.07% and 68.10%. Our findings indicate that DenseNet201 out performs ResNet50 in terms of test accuracy, demonstrating a more robust learning capability for this dataset. These results contribute to the ongoing research in deep learning-based breast cancer detection and highlight the importance of selecting an appropriate CNN architecture for medical image classification.