MobileDenseAttn:A Dual-Stream Architecture for Accurate and Interpretable Brain Tumor Detection

2025-08-22

AI Summary

To address poor generalizability, low computational efficiency, and limited interpretability in brain tumor detection from MRI scans, this paper proposes MobileDenseAttn—a dual-stream fusion model integrating MobileNetV2 and DenseNet201 to establish a lightweight-dense collaborative architecture; heterogeneous tumor representation is enhanced via feature-level fusion. Grad-CAM is incorporated for lesion localization visualization, significantly improving clinical interpretability. Leveraging data augmentation and five-fold cross-validation on 6,020 MRI volumes, the model achieves 98.35% accuracy and an F1-score of 0.9835. Compared to VGG19, it reduces inference time by 39.3% while improving accuracy by 3.67 percentage points. MobileDenseAttn thus strikes a robust balance among diagnostic accuracy, computational efficiency, and model transparency, advancing the deployment of trustworthy AI-assisted diagnosis in clinical neuroimaging.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Hardware-aware ML
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

The detection of brain tumor in MRI is an important aspect of ensuring timely diagnostics and treatment; however, manual analysis is commonly long and error-prone. Current approaches are not universal because they have limited generalization to heterogeneous tumors, are computationally inefficient, are not interpretable, and lack transparency, thus limiting trustworthiness. To overcome these issues, we introduce MobileDenseAttn, a fusion model of dual streams of MobileNetV2 and DenseNet201 that can help gradually improve the feature representation scale, computing efficiency, and visual explanations via GradCAM. Our model uses feature level fusion and is trained on an augmented dataset of 6,020 MRI scans representing glioma, meningioma, pituitary tumors, and normal samples. Measured under strict 5-fold cross-validation protocols, MobileDenseAttn provides a training accuracy of 99.75%, a testing accuracy of 98.35%, and a stable F1 score of 0.9835 (95% CI: 0.9743 to 0.9920). The extensive validation shows the stability of the model, and the comparative analysis proves that it is a great advancement over the baseline models (VGG19, DenseNet201, MobileNetV2) with a +3.67% accuracy increase and a 39.3% decrease in training time compared to VGG19. The GradCAM heatmaps clearly show tumor-affected areas, offering clinically significant localization and improving interpretability. These findings position MobileDenseAttn as an efficient, high performance, interpretable model with a high probability of becoming a clinically practical tool in identifying brain tumors in the real world.

Problem

• Detecting brain tumors in MRI scans accurately
• Improving model generalization across heterogeneous tumor types
• Enhancing computational efficiency and interpretability for clinical trust

Innovation

• Dual-stream fusion of MobileNetV2 and DenseNet201
• Feature level fusion with enhanced representation scale
• GradCAM visualization for interpretable tumor localization