Automatic glioma segmentation based on efficient U-net model using MRI images

Abstract

Gliomas are among the most aggressive and challenging brain tumors to diagnose and treat. Accurate segmentation of glioma regions in Magnetic Resonance Imaging (MRI) is essential for early diagnosis and effective treatment planning. This study proposes an optimized U-Net model tailored for glioma segmentation, addressing key challenges such as boundary delineation, computational efficiency, and generalizability. The proposed model integrates streamlined encoder-decoder pathways and optimized skip connections, achieving precise segmentation while reducing computational complexity. The model was validated on two datasets: TCGA-TCIA, containing 110 patients, and the multi-modal BraTS 2021 dataset. Comparative evaluations were conducted against state-of-the-art methods, including Attention U-Net, Trans-U-Net, DeepLabV3+, and 3D U-Net, using metrics such as Dice Coefficient, Intersection over Union (IoU), Hausdorff Distance (HD), and Structural Similarity Index (SSIM). The proposed U-Net achieved the highest performance across all metrics, with a Dice score of 92.54 %, IoU of 90.42 %, HD of 4.12 mm, and SSIM of 0.962 on the TCGA-TCIA dataset. On the BraTS dataset, it achieved comparable results, with a Dice score of 91.32 % and an IoU of 89.56 %. In contrast, other methods, such as Attention U-Net and DeepLabV3+, showed lower Dice scores of 85.62 % and 84.10 %, respectively, and higher HD values, indicating inferior boundary delineation. Additionally, the proposed model demonstrated computational efficiency, processing images in 1.5 s on average, compared to 5.0 s for Attention U-Net and 9.0 s for Trans-U-Net. These results underscore the potential of the optimized U-Net as a robust, accurate, and efficient tool for glioma segmentation. Future work will focus on clinical validation and extending the model to include automated glioma grading, further enhancing its applicability in medical imaging workflows.