ExplainabilityTransparencySafetycs.CV
An Explainable AI-Driven Framework for Automated Brain Tumor Segmentation Using an Attention-Enhanced U-Net
MD Rashidul Islam, Bakary Gibba
March 24, 2026
arXiv.org
1 citations
Key Takeaways
-
Automated AI can improve brain tumor diagnosis reliability: The model achieved 99.19% accuracy in segmenting brain tumors from MRI scans, addressing a critical clinical need where manual segmentation is time-consuming and inconsistent—potentially supporting more reliable treatment planning and patient monitoring.
-
Explainability is built into this medical AI system: The researchers used Grad-CAM visualization techniques to show which regions of brain scans the AI focuses on when making decisions, making the model's reasoning transparent to doctors rather than operating as a "black box"—an important safeguard for clinical adoption.
-
Handling data imbalance is essential for fair AI in medicine: The team used specialized loss functions (Dice Loss and Categorical Dice Loss) to prevent the model from being biased toward common tumor types, ensuring accurate detection of rarer but equally important tumor sub-regions like necrotic cores.
-
Multiple evaluation metrics reveal different aspects of AI performance: Beyond simple accuracy (99.19%), the researchers measured sensitivity and specificity separately, recognizing that missing a tumor (low sensitivity) has different consequences than false alarms (low specificity)—a nuanced approach necessary for responsible medical AI deployment.
Abstract
Computer-aided segmentation of brain tumors from MRI data is of crucial significance to clinical decision-making in diagnosis, treatment planning, and follow-up disease monitoring. Gliomas, owing to their high malignancy and heterogeneity, represent a very challenging task for accurate and reliable segmentation into intra-tumoral sub-regions. Manual segmentation is typically time-consuming and not reliable, which justifies the need for robust automated techniques.This research resolves this problem by leveraging the BraTS 2020 dataset, where we have labeled MRI scans of glioma patients with four significant classes: background/healthy tissue, necrotic/non-enhancing core, edema, and enhancing tumor. In this work, we present a new segmentation technique based on a U-Net model augmented with executed attention gates to focus on the most significant regions of images. To counter class imbalance, we employ manually designed loss functions like Dice Loss and Categorical Dice Loss, in conjunction with standard categorical cross-entropy. Other evaluation metrics, like sensitivity and specificity, were used to measure discriminability of the model between tumor classes. Besides, we introduce Grad-CAM-based explainable AI to enable visualizing attention regions and improve model interpretability, together with a smooth heatmap generation technique through Gaussian filtering. Our approach achieved superior performance with accuracy of 0.9919, Dice coefficient of 0.9901, mean IoU of 0.9873, sensitivity of 0.9908, and specificity of 0.9974. This study demonstrates that the use of attention mechanisms, personalized loss functions, and explainable AI significantly improves highly complex tumor structure segmentation precision in MRI scans, providing a reliable and explainable method for clinical applications.