Abstract
<title>Abstract</title> <p>Breast cancer remains a major cause of mortality among women, and early detection is essential for improving treatment outcomes. A mammogram is a low-dose X-ray image of the breast used to detect abnormalities, including early signs of breast cancer before symptoms become apparent. Although artificial intelligence (AI) has enhanced mammogram analysis but many existing models trained on foreign datasets that may not adequately represent local clinical environment. This study proposes an ensemble deep learning model for classifying mammogram images into Breast Imaging Reporting and Data System (BI-RADS) categories using locally sourced data. Mammogram images obtained from Total Diagnostic Center, Ibadan, Nigeria, were preprocessed to improve image quality and ensure consistency. Two deep learning models, InceptionResNetV2 and ResNet-18, were developed and combined using an ensemble approach to classify the images into BI-RADS categories. The performance of the ensemble model was evaluated using accuracy, precision, recall, and F1-score. The experimental results show that the proposed ensemble model achieved an accuracy of 99.26%, a precision of 99.27%, a recall of 99.25%, and an F1-score of 99.25%. The ensemble model also outperformed the individual InceptionResNetV2 and ResNet-18 models, demonstrating its effectiveness in accurately classifying mammogram images into BI-RADS categories. The study demonstrates the importance of developing AI models using locally representative datasets.</p>