<![CDATA[Accurate and explainable classification of brain Magnetic Resonance Imaging (MRI) is crucial for the early detection and treatment of brain tumors. This study introduces an enhanced deep learning framework that integrates transfer learning with Grad-CAM-based explainable Convolutional Neural Network (CNN) for tumor classification. The proposed approach utilizes a fine-tuned EfficientNet-B0 architecture with an optimized preprocessing pipeline consisting of Contrast Limited Adaptive Histogram Equalization (CLAHE), normalization, and multi-variant augmentation (rotation, flipping, and zoom). The model was trained on a publicly available brain MRI dataset comprising 3,000 images classified into four categories: glioma, meningioma, pituitary tumor, and non-tumor. Evaluation metrics include accuracy, precision, recall, F1-score, and AUC. Experimental results demonstrate that the proposed model achieves an accuracy of 94.2% and an AUC of 0.965, outperforming baseline CNN models by a significant margin. The use of Grad-CAM visualization provides interpretability by localizing tumor regions within MRI scans, thereby increasing the model’s clinical transparency. This study highlights the potential of explainable deep learning models to enhance diagnostic reliability in automated brain tumor detection systems.]]>
Copyrights © 2025
