Deep Learning for Benign/Malignant Classification of Breast Cancer Histopathological Images

Abstract

Early detection of breast cancer is essential to improve patient outcomes. This work proposes an automatic detection system using convolutional neural networks (CNNs) to classify histopathological images as benign or malignant. The model, trained on a large dataset, incorporates regularization techniques like dropout and data augmentation to enhance robustness. Results show high performance with an accuracy of 92.14%, recall of 91.08%, and <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$F 1$</tex>-score of 91.61%, confirming the model's potential for clinical integration.