Revolutionizing brain tumor diagnoses: a ResNet18 and focal loss approach to magnetic resonance imaging-based classification in neuro-oncology

Abstract

Brain tumor diagnosis remains a critical challenge in neuro- oncology, where accurate and timely identification of malignancies can significantly impact patient outcomes. This research explores the integration of deep learning techniques, specifically leveraging the ResNet18 architecture coupled with focal loss, to enhance the classification accuracy of magnetic resonance imaging (MRI)-based brain tumor diagnoses. ResNet18, known for its powerful feature extraction capabilities, was employed to analyze MRI scans, while focal loss was utilized to address class imbalance issues prevalent in medical datasets. The model was trained on a comprehensive dataset, achieving an accuracy of 95.54%. These results demonstrate the potential of this approach in providing robust and precise diagnostic support in clinical settings, potentially revolutionizing the current methodologies in brain tumor detection and classification. The integration of advanced neural networks with specialized loss functions presents a significant advancement in the field, paving the way for more reliable and automated neuro-oncological diagnostics.