EPRA U-Net: an efficient pyramid residual attention framework for accurate infarct segmentation in diffusion-weighted MRI.
7 affiliations
- 1Department of Computer Engineering, Yozgat Bozok University, Yozgat, Turkey. [email protected].
- 2Department of Computer Engineering, Izmir Bakircay University, Izmir, Turkey.
- 3Queen Mary's Digital Environment Research Institute (DERI), London, UK.
- 4Department of Radiology, Faculty of Medicine, Yozgat Bozok University, Yozgat, Turkey.
- 5Department of Computer Engineering, Yozgat Bozok University, Yozgat, Turkey.
- 6Department of Digital Forensics Engineering, Technology Faculty, Firat University, Elazig, Turkey.
- 7Department of Electrical Engineering, Qatar University, Doha, Qatar.
Abstract
Objective
Accurate identification of acute ischemic infarcts on diffusion-weighted magnetic resonance imaging (DWI) is a critical prerequisite for reliable lesion quantification and effective clinical decision support in the management of cerebrovascular events.Methods
This study presents EPRA U-Net (Efficient Pyramid Residual Attention U-Net), a task-specific integrated architecture for efficient and accurate infarct segmentation of DWI images. In the proposed architecture, an EfficientNet-based encoder was used as a hierarchical feature extractor with a minimized parameterization. In addition, a Residual-Recurrent (R2) block (recurrent unrolling step t = 2, following the original formulation) and Atrous Spatial Pyramid Pooling (ASPP) were integrated to enhance the performance of spatial dependency modeling. Additionally, a dual attention mechanism was incorporated to highlight lesion-related activations while concurrently enabling the suppression of extraneous background responses. To prioritize lesion detection consistent with clinical imperative, a Tversky loss function (α = 0.4, β = 0.6) was adopted, thereby emphasizing the sensitivity of detection over its specificity during the optimization process.Results
Experimental evaluations were conducted utilizing an in-house dataset comprising 167 patients with 4,895 DWI slices; subsequently, the performance of the proposed EPRA U-Net was assessed in comparison with state-of-the-art models, specifically UNet++, DeepLabV3+, and TransUNet. The experimental results suggest that EPRA U-Net attained superior performance, evidenced by a pixel-aggregated Dice of 0.8984, a per-sample Dice of 0.9469, an IoU of 0.8155, a Recall of 0.8887, a Lesion F1 of 0.9378, and an HD95 of 11.62 px (median: 2.00 px). Furthermore, a clear reduction in the rate of missed lesions, specifically by 16%, 25%, and 29%, was observed when compared with UNet++, DeepLabV3+, and TransUNet, respectively. Additionally, a comprehensive ablation study was systematically conducted.Conclusion
These findings suggest that the proposed architecture affords a more robust and accurate infarct segmentation, consequently facilitating a more reliable estimation of treatment eligibility.
Related
- AMF-U-Net: an adaptive multimodal fusion residual attention 3D U-Net for boundary-aware brain tumour segmentation
- TiLCA-Net: nuclei segmentation and classification in pathological images via task-specialized attention and efficient decoding
- Asymmetric encoder-decoder U-net with hybrid attention and residual dilated convolution for brain tumor segmentation
- A Hybrid Residual-Swin Transformer Design with Attention for Prostate Cancer Segmentation
- Brain tumor segmentation using a combination of graph cut and optimized U-Net with genetic algorithm: An efficient hybrid framework for enhanced medical image analysis
BibTeX
@article{ulutas2026epra,
title = {{EPRA U-Net: an efficient pyramid residual attention framework for accurate infarct segmentation in diffusion-weighted MRI.}},
author = {Hasan Ulutas and Muhammet Emin Sahin and Mustafa Fatih Erkoc and Esra Yuce and Turker Tuncer and Sengul Dogan and Serkan Kiranyaz},
journal = {BMC medical imaging},
year = {2026},
month = jul,
doi = {10.1186/s12880-026-02569-9},
url = {https://europepmc.org/article/MED/42919228},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - EPRA U-Net: an efficient pyramid residual attention framework for accurate infarct segmentation in diffusion-weighted MRI.
AU - Hasan Ulutas
AU - Muhammet Emin Sahin
AU - Mustafa Fatih Erkoc
AU - Esra Yuce
AU - Turker Tuncer
AU - Sengul Dogan
AU - Serkan Kiranyaz
PY - 2026
DA - 2026/07/01
JO - BMC medical imaging
DO - 10.1186/s12880-026-02569-9
UR - https://europepmc.org/article/MED/42919228
AB - Objective Accurate identification of acute ischemic infarcts on diffusion-weighted magnetic resonance imaging (DWI) is a critical prerequisite for reliable lesion quantification and effective clinical decision support in the management of cerebrovascular events.Methods This study presents EPRA U-Net (Efficient Pyramid Residual Attention U-Net), a task-specific integrated architecture for efficient and accurate infarct segmentation of DWI images. In the proposed architecture, an EfficientNet-based encoder was used as a hierarchical feature extractor with a minimized parameterization. In addition, a Residual-Recurrent (R2) block (recurrent unrolling step t = 2, following the original formulation) and Atrous Spatial Pyramid Pooling (ASPP) were integrated to enhance the performance of spatial dependency modeling. Additionally, a dual attention mechanism was incorporated to highlight lesion-related activations while concurrently enabling the suppression of extraneous background responses. To prioritize lesion detection consistent with clinical imperative, a Tversky loss function (α = 0.4, β = 0.6) was adopted, thereby emphasizing the sensitivity of detection over its specificity during the optimization process.Results Experimental evaluations were conducted utilizing an in-house dataset comprising 167 patients with 4,895 DWI slices; subsequently, the performance of the proposed EPRA U-Net was assessed in comparison with state-of-the-art models, specifically UNet++, DeepLabV3+, and TransUNet. The experimental results suggest that EPRA U-Net attained superior performance, evidenced by a pixel-aggregated Dice of 0.8984, a per-sample Dice of 0.9469, an IoU of 0.8155, a Recall of 0.8887, a Lesion F1 of 0.9378, and an HD95 of 11.62 px (median: 2.00 px). Furthermore, a clear reduction in the rate of missed lesions, specifically by 16%, 25%, and 29%, was observed when compared with UNet++, DeepLabV3+, and TransUNet, respectively. Additionally, a comprehensive ablation study was systematically conducted.Conclusion These findings suggest that the proposed architecture affords a more robust and accurate infarct segmentation, consequently facilitating a more reliable estimation of treatment eligibility.
SN - 1471-2342
ER -