Brain tumor classification from magnetic resonance imaging (MRI) is important for timely diagnosis and treatment planning. This study proposes a hybrid deep learning framework for four-class brain tumor MRI classification using ResNet50 as a deep feature extractor and a Multi-Layer Perceptron (MLP) as the final classifier. The framework was evaluated on the publicly available Kaggle Brain Tumor MRI Dataset containing 7,023 images across glioma, meningioma, pituitary tumor, and no-tumor classes. The original test set of 1,311 images was retained exclusively for final evaluation. The proposed ResNet50 + MLP model achieved 97.64% accuracy, 97.61% precision, 97.58% recall, 97.59% F1-score, 99.21% specificity, and 99.61% AUC, outperforming CNN, VGG16, InceptionV3, DenseNet121, EfficientNetB0, MobileNetV3, and standalone ResNet50 under identical experimental conditions. Grad-CAM visualizations were used to provide visual interpretability of model predictions. The results demonstrate that the proposed framework offers accurate and interpretable support for automated brain tumor MRI classification.
This study presented a hybrid ResNet50–MLP framework for automated multi-class brain tumor classification using magnetic resonance imaging (MRI). The proposed framework integrates transfer learning-based deep feature extraction through a pretrained ResNet50 backbone with a lightweight Multi-Layer Perceptron (MLP) classifier to accurately distinguish four clinically important categories: glioma, meningioma, pituitary tumor, and no-tumor cases. A comprehensive preprocessing pipeline, including image resizing, normalization, contrast enhancement, skull stripping, Fuzzy C-Means (FCM)-based region enhancement, and data augmentation, was employed to improve image quality and enhance the discriminative capability of the extracted features. Furthermore, transfer learning enabled the model to leverage knowledge learned from large-scale natural image datasets, resulting in improved convergence, better feature representation, and enhanced classification performance on a relatively limited medical imaging dataset.
The proposed framework was evaluated on the publicly available Kaggle Brain Tumor MRI Dataset consisting of 7,023 contrast-enhanced T1-weighted MRI images. Experimental results demonstrated that the proposed model consistently outperformed several widely used deep learning architectures, including CNN, VGG16, InceptionV3, DenseNet121, EfficientNetB0, standalone ResNet50, and MobileNetV3, under identical preprocessing, training, and evaluation conditions. The proposed framework achieved a test accuracy of 97.64%, precision of 97.61%, recall of 97.58%, F1-score of 97.59%, specificity of 99.21%, AUC of 99.61%, and Cohen’s Kappa coefficient of 0.968, demonstrating its effectiveness for reliable multi-class brain tumor classification. Class-wise analysis further confirmed balanced performance across all tumor categories, while Grad-CAM visualizations provided meaningful explanations by highlighting clinically relevant tumor regions, thereby improving the transparency and interpretability of the proposed framework for potential clinical use.
Despite these encouraging results, several limitations should be acknowledged. The experimental evaluation was performed using a single publicly available dataset, which may not fully represent the diversity of MRI acquisition protocols, scanners, and patient populations encountered in real-world clinical practice. In addition, the proposed framework operates on two-dimensional MRI slices rather than complete three-dimensional volumetric scans, limiting its ability to exploit inter-slice spatial information. Moreover, patient-wise cross-validation and external multi-center validation could not be performed because of dataset constraints, which may affect the assessment of the model’s generalization capability across different clinical environments [5, 12, 16].
References
1. Abdulsattar, N.S., Abdulsattar, F.S.: Brain-tumor classification using resnet50 enhanced with se and cbam attention mechanisms. Jordanian Journal of Computers and Information Technology (JJCIT) 12(1), 98–111 (2026). https://doi.org/10. 5455/jjcit.71-1764868722
2. Aboussaleh, I., Riffi, J., Mahraz, A.M., Tairi, H.: Brain tumor segmentation based on deep learning’s feature representation. Journal of Imaging 7(12), 269 (2021)
3. Agarwal, M., Rani, G., Kumar, A., Kumar, P., Manikandan, R., Gandomi, A.H.: Deep learning for enhanced brain tumor detection and classification. Results in Engineering 22, 102117 (2024)
4. Asif, R.N., Naseem, M.T., Ahmad, M., Mazhar, T., Khan, M.A., Khan, M.A.,Al-Rasheed, A., Hamam, H.: Brain tumor detection empowered with ensemble deep learning approaches from mri scan images. Scientific reports 15(1), 15002 (2025)
5. Chowdhury, M.F.K., Ferdous, J.: Brain tumor classification in mri images: A computationally efficient convolutional neural network. In: 2025 IEEE International Conference on Biomedical Engineering, Computer and Information Technology for Health (BECITHCON). pp. 633–638. IEEE (2025)
6. Dheepak, G., Vaishali, D., et al.: Brain tumor classification: a novel approach integrating glcm, lbp and composite features. Frontiers in oncology 13, 1248452 (2024)
7. Ghasemi, R., Islam, N., Bayat, S., Shabir, M., Rahman, S., Amin, F., de la Torre, I., Castilla, A´.K., Garc´ıa, D.L.R.V.: Detection and classification of brain tumor using a hybrid learning model in ct scan images. Scientific Reports 15(1), 35085 (2025)
8. Guluwadi, S., et al.: Enhancing brain tumor detection in mri images through explainable ai using grad-cam with resnet 50. BMC medical imaging 24(1), 1–19 (2024)
9. Hosny, K.M., Mohammed, M.A.: Explainable ai and vision transformers for detection and classification of brain tumor: a comprehensive survey. Artificial Intelligence Review 58(9), 259 (2025)
10. Iftikhar, S., Anjum, N., Siddiqui, A.B., Ur Rehman, M., Ramzan, N.: Explainable cnn for brain tumor detection and classification through xai based key features identification. Brain Informatics 12(1), 10 (2025)
11. JS, N., et al.: Brain tumor segmentation using multi-scale attention u-net with efficientnetb4 encoder for enhanced mri analysis. Scientific Reports 15(1), 1–20 (2025)
12. Jyothi, P., Singh, A.R.: Deep learning models and traditional automated techniques for brain tumor segmentation in mri: a review. Artificial intelligence review 56(4), 2923–2969 (2023)
13. Kaplan, K., Kaya, Y., Kuncan, M., Ertunc¸, H.M.: Brain tumor classification using modified local binary patterns (lbp) feature extraction methods. Medical hypotheses 139, 109696 (2020)
14. Khan, S.U.R., Zhao, M., Li, Y.: Detection of mri brain tumor using residual skip block based modified mobilenet model. Cluster Computing 28(4), 248 (2025)
15. Koc, A.B., Akgun, D.: A review of u-net based deep learning frameworks for mri-based brain tumor segmentation. Diagnostics 16(4), 506 (2026)
16. Li, Z., Dib, O.: Empowering brain tumor diagnosis through explainable deep learning. Machine Learning and Knowledge Extraction 6(4), 2248–2281 (2024)
17. Maqsood, R., Abid, F., Rasheed, J., Osman, O., Alsubai, S.: Optimal res-unet architecture with deep supervision for tumor segmentation. Frontiers in Medicine 12, 1593016 (2025)
18. Mehrdad, V., Talebzadeh, R., Fazaeli, N.: Brain tumor classification using optimized resnet50 with dynamic precision optimization for enhanced speed and diagnostic accuracy. Scientific Reports (2026)
19. Mohan, P., Ramkumar, G.: An inception v3 based glioma brain tumor detection in mri images. In: 2024 5th International Conference on Circuits, Control, Communication and Computing (I4C). pp. 560–564. IEEE (2024)
20. Napa, K.K., Murugan, S., Murugan, J.S., Jayanthi, A.: Brain-meta: A reproducible cnn–vision transformer meta-ensemble pipeline for explainable brain tumor classifi-cation. MethodsX 16, 103769 (2026)
21. Nayak, D.R., Padhy, N., Mallick, P.K., Zymbler, M., Kumar, S.: Brain tumor classification using dense efficient-net. Axioms 11(1), 34 (2022)
22. Nickparvar, M.: Brain tumor mri dataset (2021), https://www.kaggle.com/ datasets/masoudnickparvar/brain-tumor-mri-dataset, accessed: 24 May 2026
23. Pillai, R., Sharma, A., Sharma, N., Gupta, R.: Brain tumor classification using vgg 16, resnet50, and inception v3 transfer learning models. In: 2023 2nd International Conference for Innovation in Technology (INOCON). pp. 1–5. IEEE (2023)
24. Prabhas, K.S., Basem, A., Lakshmi, L., Talha, A., Mohammed, S.H., Khan, M.I., Khedher, N.B.: A deep learning framework for brain tumor detection using cnns and transfer learning on mri scans. Systems and Soft Computing p. 200389 (2025)
25. Putri, F.A.P.R., Tanjung, J.P., Dharshini, N.P.: Implementation of transfer learning on cnn using densenet121 and resnet50 for brain tumor classification. Journal of Informatics and Telecommunication Engineering 8(2) (2025). https://doi.org/ 10.31289/jite.v8i2.13952
26. Rahman, A., Hayat, M., Iqbal, N., Alarfaj, F.K., Alkhalaf, S., Alturise, F.: Enhanced mri brain tumor detection using deep learning in conjunction with explainable ai shap based diverse and multi feature analysis. Scientific Reports 15(1), 29411 (2025)
27. Rahman, T., Islam, M.S.: Mri brain tumor detection and classification using parallel deep convolutional neural networks. Measurement: Sensors 26, 100694 (2023)
28. Rajendran, S., Rajagopal, S.K., Thanarajan, T., Shankar, K., Kumar, S., Alsubaie, N.M., Ishak, M.K., Mostafa, S.M.: Automated segmentation of brain tumor mri images using deep learning. IEEE Access 11, 64758–64768 (2023)
How to Cite This Paper
Riyanshu Saini, Sumika Jain, Varun Bansal (2026). Interpretable Framework for Brain Tumor Classification from MRI^. International Journal of Computer Techniques, 13(4). ISSN: 2394-2231.
