Ensembling Transformer-Based Models for 3D Ischemic Stroke Segmentation in Non-Contrast CT
Source abstract
Abstract
Ischemic stroke remains one of the leading causes of mortality and disability, and accurate segmentation of the affected areas on CT brain images plays a crucial role in timely diagnosis and clinical decision-making. This study proposes an ensemble approach based on the combination of the transformer-based models SE-UNETR and Swin UNETR using a weighted voting strategy. Its performance was evaluated using the Dice similarity coefficient, which quantifies the overlap between the predicted lesion regions and the ground-truth annotations. In this study, three-dimensional CT scans of the brain from 98 patients with a confirmed diagnosis of acute ischemic stroke were used. The data were provided by the International Tomography Center, SB RAS. The experimental results demonstrated that the ensemble based on transformer models significantly outperforms each individual model, providing more stable and accurate predictions. The final Dice coefficient reached 0.7983, indicating the high effectiveness of the proposed approach for ischemic lesion segmentation in CT images. The analysis showed more precise delineation of ischemic lesion boundaries and a reduction in segmentation errors. The proposed method can serve as an effective tool in automated stroke diagnosis systems and other applications requiring high-accuracy medical image analysis.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Medical Image Segmentation Techniques · Medical Imaging and Analysis · Brain Tumor Detection and Classification
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Department of Computer Science, Al-Farabi Kazakh National University, Almaty 050040, Kazakhstan
Department of Mechanics and Mathematics, Novosibirsk State University, 630090 Novosibirsk, Russia
Sobolev Institute of Mathematics SB RAS, 630090 Novosibirsk, Russia
Higher School of Information Technology and Engineering, Astana International University, Astana 010000, Kazakhstan
Department of Computer Engineering, Astana IT University, Astana 010000, Kazakhstan
Funding
Committee of Science of the Ministry of Science and Higher Education of the Republic of Kazakhstan
References (37)
Show source-provided references
1. Bakator, M., and Radosav, D. (2018). Deep Learning and Medical Diagnosis: A Review of Literature. Multimodal Technol. Interact., 2.10.3390/mti2030047 ↗
2. Zhu, S., Xia, X., Zhang, Q., and Belloulata, K. (2007, January 16–18). An Image Segmentation Algorithm in Image Processing Based on Threshold Segmentation. Proceedings of the 2007 Third International IEEE Conference on Signal-Image Technologies and Internet-Based System, Shanghai, China.10.1109/SITIS.2007.116 ↗
3. Chen, X., Williams, B.M., Vallabhaneni, S.R., Czanner, G., Williams, R., and Zheng, Y. (2019, January 16–20). Learning Active Contour Models for Medical Image Segmentation. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA.10.1109/CVPR.2019.01190 ↗
4. Deep Neural Architectures for Medical Image Semantic Segmentation: Review10.1109/ACCESS.2021.3086530 ↗
5. Tursynova, A., and Omarov, B. (2021, January 25–26). 3D U-Net for brain stroke lesion segmentation on ISLES 2018 dataset. Proceedings of the 2021 16th International Conference on Electronics Computer and Computation (ICECCO), Kaskelen, Kazakhstan.10.1109/ICECCO53203.2021.9663825 ↗
6. A hybrid attention-based residual Unet for semantic segmentation of brain tumor
7. Zhang, B., Qiu, S., and Liang, T. (2024). Dual Attention-Based 3D U-Net Liver Segmentation Algorithm on CT Images. Bioengineering, 11.10.3390/bioengineering11070737 ↗
8. U-Net and Its Variants for Medical Image Segmentation: A Review of Theory and Applications10.1109/ACCESS.2021.3086020 ↗
9. Pinheiro, G.R., Voltoline, R., Bento, M., and Rittner, L. (2019). V-Net and U-Net for ischemic stroke lesion segmentation in a small dataset of perfusion data. Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries (BrainLes 2018), Springer.10.1007/978-3-030-11723-8_30 ↗
10. Atika, L., Nurmaini, S., Partan, R.U., and Sukandi, E. (2022). Image segmentation for mitral regurgitation with convolutional neural network based on UNet, ResNet, VNet, FractalNet, and SegNet: A preliminary study. Big Data Cogn. Comput., 6.10.3390/bdcc6040141 ↗
11. Unified focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation10.1016/j.compmedimag.2021.102026 ↗
12. Garcia-Salgado, B.P., Almaraz-Damian, J.A., Cervantes-Chavarria, O., Ponomaryov, V., Reyes-Reyes, R., Cruz-Ramos, C., and Sadovnychiy, S. (2024). Enhanced Ischemic Stroke Lesion Segmentation in MRI Using Attention U-Net with Generalized Dice Focal Loss. Appl. Sci., 14.10.3390/app14188183 ↗
13. Deep convolutional neural network for automatically segmenting acute ischemic stroke lesion in multi-modality MRI10.1007/s00521-019-04096-x ↗
14. Amirgaliyev, Y.N., Buribayev, Z.A., Melis, Z.M., and Ataniyazova, A.S. (2021, January 19–22). On one approach to recognizing fuzzy images of faces based on an ensemble. Proceedings of the 25th International Conference on Circuits, Systems, Communications and Computers (CSCC 2021), Crete Island, Greece.10.1109/CSCC53858.2021.00011 ↗
15. Searching for optimal classifier using a combination of cluster ensemble and kernel method
16. Acute ischemic stroke lesion segmentation in non-contrast CT images using 3D convolutional neural networks10.18287/2412-6179-CO-1233 ↗
17. Vezakis, A., Vezakis, I., Vagenas, T.P., Kakkos, I., and Matsopoulos, G.K. (2024). A Multidimensional Framework Incorporating 2D U-Net and 3D Attention U-Net for the Segmentation of Organs from 3D Fluorodeoxyglucose-Positron Emission Tomography Images. Electronics, 13.10.3390/electronics13173526 ↗
18. Yousef, R., Khan, S., Gupta, G., Siddiqui, T., Albahlal, B.M., Alajlan, S.A., and Haq, M.A. (2023). U-Net-Based Models towards Optimal MR Brain Image Segmentation. Diagnostics, 13.10.3390/diagnostics13091624 ↗
19. Zhang, L., Deng, X., and Lu, Y. (2023, January 5–8). Segment Anything Model (SAM) for Medical Image Segmentation: A Preliminary Review. Proceedings of the 2023 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Istanbul, Turkiye.10.1109/BIBM58861.2023.10386032 ↗
20. Hatamizadeh, A., Tang, Y., Nath, V., Yang, D., Myronenko, A., Landman, B., Roth, H., and Xu, D. (2022, January 3–8). UNETR: Transformers for 3D Medical Image Segmentation. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Waikoloa, HI, USA.10.1109/WACV51458.2022.00181 ↗
21. Hatamizadeh, A., Yang, D., Roth, H.R., and Xu, D. (2022). Swin UNETR: Swin Transformers for Semantic Segmentation of Brain Tumors in MRI Images. arXiv.10.1007/978-3-031-08999-2_22 ↗
22. Tang, Y., Yang, D., Li, W., Roth, H.R., Landman, B., and Xu, D. (2022). Self-supervised Pre-training of Swin Transformers for 3D Medical Image Analysis. arXiv.10.1109/CVPR52688.2022.02007 ↗
23. Wang, Y., Li, Z., Mei, J., Wei, Z., Liu, L., Wang, C., Sang, S., Yuille, A., Xie, C., and Zhou, Y. (2023). SwinMM: Masked Multi-view with Swin Transformers for 3D Medical Image Segmentation. arXiv.10.1007/978-3-031-43898-1_47 ↗
24. Jabal, M.S., Joly, O., Kallmes, D., Harston, G., Rabinstein, A., Huynh, T., and Brinjikji, W. (2022). Interpretable machine learning modeling for ischemic stroke outcome prediction. Front. Neurol., 13.10.3389/fneur.2022.884693 ↗
25. Tynymbaev, S., Chinibaeva, T.T., Temirbekova, Z.E., and Begimbayeva, Y.Y. (2024). Device Multiplication by Two Polynomials Modulo Irreducible Polynomials. (No. 36794), Patent RK.
26. An medical image file formats and digital image conversion10.35940/ijeat.A1093.1291S419 ↗
27. Brain CT image segmentation based on 3D slicer10.21595/chs.2020.21263 ↗
28. Deep MRI brain extraction: A 3D convolutional neural network for skull stripping10.1016/j.neuroimage.2016.01.024 ↗
29. A large comparison of normalization methods on time series10.1016/j.bdr.2023.100407 ↗
30. Delving Deep into Spatial Pooling for Squeeze and Excitation Networks10.1016/j.patcog.2021.108159 ↗
31. Taha, A.A., and Hanbury, A. (2015). Metrics for evaluating 3D medical image segmentation: Analysis, selection, and tool. BMC Med. Imaging, 15.10.1186/s12880-015-0068-x ↗
32. Evaluation: From precision, recall and F-measure to ROC, informedness, markedness & correlation
33. Ogundokun, R.O., Maskeliunas, R., Misra, S., and Damaševičius, R. (2022). Improved CNN based on batch normalization and Adam optimizer. Computational Science and Its Applications, Proceedings of the International Conference on Computational Science and Its Applications, Malaga, Spain, 4–7 July 2022, Springer International Publishing.10.1007/978-3-031-10548-7_43 ↗
34. NVIDIA A100 Tensor Core GPU: Performance and innovation10.1109/MM.2021.3061394 ↗
35. METrans: Multi-encoder Transformer for Ischemic Stroke Segmentation
36. Omarov, M., Ibragimov, M., Kaldybekov, K., and Aytmahanov, M. (2022). Modified 3D U-Net for Brain Stroke Lesion Segmentation on Computed Tomography Images. Computers, 11.10.32604/cmc.2022.020998 ↗
37. Checklist for Artificial Intelligence in Medical Imaging (CLAIM): 2024 Update10.1148/ryai.240300 ↗