Aug 2026· Artificial Intelligence and Applications· 0 citations· 41 references
TL;DR
A unique explainable deep learning model that integrates Tiny-ConvNeXt and DenseNet169 using multi-stage brain tumor classification is proposed, indicating that the suggested model provides a transparent and reliable framework for brain tumor detection, with potential applications in practical clinical decision support systems.
Abstract
Early and accurate detection of brain tumors remains a major challenge in medical imaging due to limited dataset size, trained deep learning models, patient variability, and the complexity of manual interpretation; traditional approaches sometimes rely on lower-level feature extraction, which may not apply well to medical images. To address these challenges, we propose a unique explainable deep learning model that integrates Tiny-ConvNeXt and DenseNet169 using multi-stage brain tumor classification. Two clinically approved, open-access Magnetic Resonance Imaging (MRI) datasets were combined to generate a bigger, more customized dataset, and substantial data augmentation techniques were used to improve model generalization. According to experimental results, our proposed model outperforms current state-of-the-art methods with a classification accuracy of 99.74%. Additionally, statistical significance tests confirmed the incremental contribution of each model component, and ablation studies showed the robustness of the conclusions. Additionally, explainable artificial intelligence techniques like Local Interpretable Model-agnostic Explanations and Gradient-weighted Class Activation Mapping++ were employed to enhance interpretability, enabling visual explanations of tumor localization. These results indicate that the suggested model provides a transparent and reliable framework for brain tumor detection, with potential applications in practical clinical decision support systems.
Received: 18 May 2025 | Revised: 29 April 2026 | Accepted: 7 July 2026
Conflicts of Interest
The authors declare that they have no conflicts of interest to this work.
Data Availability Statement
The data that support the findings of this study are openly available in the American Society of Clinical Oncology at https://www.cancer.org/cancer/types/brain-spinal-cord-tumors-adults/key-statistics.html, in Kaggle at https://www.kaggle.com/datasets/ahmedhamada0/brain-tumor-detection, and in Kaggle at https://www.kaggle.com/datasets/navoneel/brain-mri-images-for-brain-tumor-detection.
Author Contribution Statement
Md Sadi Al Huda: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Writing - original draft, Writing - review & editing, Visualization, Supervision, Project administration. Kazi Tanvir: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Data curation, Visualization. Zubaida Akhter: Software, Validation, Formal analysis, Resources, Data curation, Writing - original draft, Writing - review & editing, Visualization. Md. Shahidul Khan Pappo: Validation, Writing - review & editing, Visualization. Md. Asraf Ali: Resources, Writing - review & editing, Supervision, Project administration. Nasim Ahmed: Resources, Writing - review & editing, Supervision, Project administration.
This study presents innovative approaches for diagnosing and classifying brain tumors from MRI images using advanced deep learning models to address
clinical data constraints
— including single-center acquisition, slice-level (not pixel-level) labeling, and real-world imaging variability — alongside extreme class imbalance. By combining convolutional neural networks (CNNs) with recurrent architectures and applying optimization algorithms like Adam and RMSprop, we improve detection accuracy and enhance interpretability. A dataset of 540 MRI images was collected from Hospital, covering various tumor types and patient backgrounds. Preprocessing steps such as normalization, segmentation, and texture-based feature extraction were applied to enhance data quality and model performance. A key innovation of this work is the use of a deep pre-trained model (VGG16), which demonstrates strong generalization potential for future clinical applications. Additionally, our use of real-world hospital data—more challenging than standard Kaggle datasets—makes our results more applicable in practice. Experimental results show that CNN-based models, especially VGG16 and ResNet v2, significantly outperform traditional methods. The VGG16 model achieved a classification accuracy of 97.1%, compared to 85.75% for Random Forest and 83.00% for Support Vector Machine. Crucially, these improvements reveal a critical trade-off: while macro-accuracy reaches 97.1%, Metastatic tumor recall remains at 75% — underscoring that clinical AI must prioritize equitable error distribution over aggregate metrics. Despite these advances, challenges remain, including imbalanced data and preprocessing limitations. Future research should focus on better data balancing and hybrid optimization strategies. This study contributes a robust framework for brain tumor detection, offering practical value for medical imaging and improved patient outcomes.
Magnetic resonance imaging is widely used for the examination of brain abnormalities because it provides detailed soft-tissue information without ionizing radiation. Nevertheless, manual interpretation of large numbers of MRI slices is time-consuming and may be affected by inter-observer variation. This paper presents a structured deep learning framework for multi-class brain tumor classification using the EfficientNet family of convolutional neural networks. The framework emphasizes consistent preprocessing, transfer learning, class-balanced augmentation, careful validation, and clinically meaningful performance reporting. Rather than treating classification as an isolated model-training task, the proposed approach connects data quality, model calibration, error analysis, and reproducibility. The paper also explains why EfficientNet is suitable for medical image classification: its compound scaling strategy balances network depth, width, and input resolution, enabling strong feature learning with comparatively efficient computation. A complete experimental protocol is described for separating training, validation, and test data at the patient level; controlling information leakage; selecting evaluation metrics; and comparing EfficientNet variants with conventional convolutional baselines. The resulting framework can support reliable thesis-level experimentation and can later be extended with visual explanation methods such as Grad-CAM. The study concludes that model efficiency alone is insufficient; dependable brain tumor classification requires disciplined data handling, transparent reporting, and external validation.
Ajay Khatri, Sanmati Jain· International Journal of Eng...· 0 citations
This study investigates the application of deep learning architectures, including Convolutional Neural Network, VGG16, VGG19, ResNet50, and MobileNet, for brain tumor detection and classification and confirms that advanced deep learning architectures not only achieve high classification accuracy but also improve interpretability, thereby offering reliable and clinically applicable solutions for automated brain tumor diagnosis.
H. Uzel, Feyyaz Alpsalaz, Yıldırım Özüpak et al.· Computers and Electronics in...· 0 citations
Brain tumors represent a significant clinical challenge, with accurate classification being essential for treatment planning. Current deep learning approaches face two critical limitations: insufficient robust-ness across diverse imaging conditions and lack of clinical interpretability. Here we present an ensemble deep learning framework integrating three complementary architectures (MobileNetV2, EfficientNetB3, and DenseNet121) with spatial attention mechanisms and explainability features. Using 7,023 MRI images across four diagnostic categories (glioma, meningioma, pituitary tumor, and tumor-absent), our approach achieved 98.47% classification accuracy with balanced performance across all classes (F1-scores: 98.12% for glioma, 97.89% for meningioma, 98.76% for pituitary, 99.21% for tumor-absent cases). The ensemble demonstrated statistically significant improvement over individual models (p < 0.01, McNemar’s test), with gains of 1.24-1.75 percentage points. Integration of Gradient-weighted Class Activation Map-ping provided interpretable visual explanations with activation patterns consistently focusing on tumor regions. The findings demonstrate the potential of combining ensemble learning, attention, and visual explanation for brain tumor classification. However, the results represent internal validation on a single partition of aggregated public datasets and require confirmation through repeated validation and independent clinical evaluation.
Unknown authors· International Journal of Eng...· 0 citations
Accurate and explainable automated brain tumor classification using magnetic resonance imaging (MRI) remains an important challenge in medical image analysis. Transfer-learning studies often report high accuracy on public MRI benchmarks, but comparisons can be difficult to interpret when architectures are trained under different protocols or evaluated without statistical and explainability analyses. Here, we present an explainable transfer-learning framework for four-class brain tumor classification (glioma, meningioma, pituitary tumor, and no-tumor) in which MobileNetV2, ResNet50, and EfficientNetB0 are compared under a common training protocol. Models were trained and evaluated on the publicly available Brain Tumor MRI Dataset containing 7,200 T1-weighted contrast-enhanced axial images (5,600 training and 1,600 testing images, balanced across the four classes). The three backbones were fine-tuned using the same two-phase training strategy and compared using accuracy, precision, recall, F1-score, one-vs-rest ROC-AUC, confusion matrices, Cohen's kappa, Wilson 95% confidence intervals, McNemar paired tests, ablation analysis, and computational-cost measures. ResNet50 achieved the highest accuracy (96.06%; 95% CI: 94.99–96.91%) and macro-F1 (0.960), with a macro-averaged ROC-AUC of 0.990 and Cohen's kappa of 0.948. Its difference from EfficientNetB0 was statistically significant (McNemar, p = 0.003), whereas the difference from MobileNetV2 was not (p = 0.263). MobileNetV2 retained 99.6% of ResNet50's accuracy with substantially lower parameter and arithmetic cost. Per-class analysis identified glioma as the most difficult class. Grad-CAM was examined on correct and incorrect predictions and cross-checked with complementary attribution methods; a quantitative border-mass analysis over 200 correctly classified images per group did not support the specific hypothesis that no-tumor predictions were driven by a top-border watermark. Overall, the study supports controlled statistical, computational, and explainability analysis as a useful framework for benchmarking brain tumor MRI classifiers, while external validation and broader robustness assessment remain necessary before clinical use.
Sif K. Ebis· Journal of Al-Farabi for Eng...· 0 citations
Brain tumor classification from MRI scans is a clinically critical task, and deep learning approaches have shown strong potential in accurately detecting various types of brain tumors (BTs). This paper presents a comparative evaluation of seven deep learning architectures for BT detection: VGG-16, ResNet-18, ResNet-50, DenseNet-121, EfficientNet-B0, ConvNeXt-Tiny, and ViT-Base/16. The models are assessed using two publicly available datasets: the Kaggle brain tumor dataset (7,023 images across four classes) and the Figshare dataset (3,064 images across three classes). To ensure robust and unbiased evaluation, a leakage-aware experimental protocol is adopted, incorporating an 80/20 development–test set split along with three-fold cross-validation. When trained exclusively on the Figshare dataset, ViT-Base/16 and VGG-16 achieve relatively low accuracies of 56.12% and 46.49%, respectively. However, the application of transfer learning substantially improves their performance, increasing accuracy to 98.21% and 97.88%. On the Kaggle dataset, EfficientNet-B0 achieves the best overall performance, reaching 99.07% accuracy and 99.70% specificity. Furthermore, an analysis of model accuracy versus parameter size demonstrates that DenseNet-121 and EfficientNet-B0 require approximately four and six times fewer parameters than ConvNeXt-Tiny and ResNet-50, while maintaining competitive performance. These findings suggest that compact architectures can provide an effective balance between classification accuracy and computational efficiency.
Ahmad Fahim Faqiri, E. A. Ince· Signal Processing and Commun...· 0 citations
