Brain Tumor Detection from MRI Images Based on Deep Learning and Machine Learning Approaches

Le résumé fourni par la source

Significant progress in artificial intelligence and medical imaging has dramatically enhanced disease analysis and prediction, particularly for brain tumor identification. Magnetic Resonance Imaging (MRI) is the primary method for detecting brain tumors. Modern imaging modalities now enable comprehensive 3D brain scans, providing critical, detailed views for tumor diagnosis. A serious step in this diagnostic development comprises isolating relevant features from MRI scans, an area where several methodologies have been proposed. In this study, we introduced a novel method for detecting brain tumors in MRI scans, focusing on developing a precise system using deep learning (DL) and machine learning (ML) techniques. The methodology begins with image acquisition, followed by preprocessing stages employing an Adaptive Contrast Enhancement Algorithm (ACEA) and a median filter. Segmentation of the enhanced images is then performed via the Fuzzy c-means algorithm. Subsequently, key textural features, including energy, entropy, mean, and contrast, are derived employing the Gray-level co-occurrence matrix (GLCM). Finally, abnormal tissue categorization is achieved through the proposed hybrid enhanced convolutional neural network variant, known as AlexNet, integrated with a Support Vector Machine (AlexNet-SVM) classifier. A hybrid learning methodology integrates these technologies to further boost the network accuracy. A comparative analysis reveals that the proposed approach outperformed all traditional approaches, achieving an accuracy of 98.28%. The findings demonstrate that this integrated deep learning system delivers exceptional accuracy and precision, showing strong potential as a diagnostic tool for reliable brain tumor identification.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.