Date of Publication :15th March 2025

Abstract: Identifying brain tumors in MRI images is an essential medical imaging endeavor that helps with early diagnosis and therapy planning. Automated deep-learning-based methods are required for precise categorization. Deep neural networks have become effective models for identifying brain tumors, thereby increasing the accuracy and efficiency of medical imaging. Current sol utions rely on manual feature extraction methods and traditional machine-learning models, which frequently result in poor accuracy and misclassification. To categorize brain tumors into four groups—gliomas, meningiomas, pituitary tumors, and healthy cases—our suggested approach combines Convolutional Neural Networks (CNNs) and ResNet-50. This approach guarantees accurate tumor identification from MRI scans, with an accuracy of 98.19%. By improving the feature extraction, our method guarantees reliabl e tumor categorization. In medical imaging, this sophisticated framework greatly enhances early tumor detection, reduces diagnostic errors, and maximizes model generalization, resulting in higher diagnostic accuracy and dependability.

Reference :