Early and precise diagnosis of brain tumors is critical for effective treatment planning and patient outcomes. While MRI is the predominant diagnostic tool, manual image evaluation can be subjective and time-consuming. This study introduces a novel multi-model deep learning framework integrating a custom convolutional neural network with pre-trained VGG16 and ResNet50 architectures for classifying glioma, meningioma, pituitary tumor, and non-tumor cases from MRI data. Feature fusion and optimization techniques were employed to enhance generalization and diagnostic accuracy. The proposed ensemble model achieves a classification accuracy of 99.69% on the test dataset, demonstrating superior performance compared to the individual models in terms of precision, recall, and F1-score as well. The optimized model convergence and feature learning contribute to its robustness on the MRI dataset. This work presents a robust deep learning framework for automated brain tumor classification from MRI, offering high accuracy and potentially aiding radiologists in efficient diagnosis. |
*** Title, author list and abstract as submitted during Camera-Ready version delivery. Small changes that may have occurred during processing by Springer may not appear in this window.