Transfer Learning Models for MRI-Based Brain Tumor Detection

Human health, especially brain health, is paramount for overall well-being. Magnetic Resonance Imaging (MRI) serves as a vital diagnostic tool for assessing critical organs such as the brain. The abundance of data generated by MRI scans presents a significant opportunity for artificial intelligence (AI) to aid in health decision-making processes. In this paper, we focus on using AI techniques, particularly transfer learning, to classify brain tumors from MRI images. By leveraging transfer learning, we aim to enhance the performance of image processing algorithms in the classification of brain tumors. Our approach involves employing established architectures such as Xception, ResNet50, and DenseNet201, adapted through transfer learning for this specific task. Among the architectures tested, our results highlight the superiority of the ResNet50 model in binary classification of brain tumors using MRI scans. The proposed model achieves remarkable performance metrics, including an accuracy of $99 \%$, precision of $99.5 \%$, recall of $99.5 \%$, and F1 Score of $99 \%$. This research underscores the potential of transfer learning in advancing diagnostic capabilities for brain health assessment.