MRG'DE İSKEMİK İNME SEGMENTASYONU İÇİN DERİN ÖĞRENME MODELLERİNİN KARŞILAŞTIRMALI ANALİZİ

Öz

İnme, küresel çapta önde gelen bir ölüm ve uzun süreli sakatlık nedenidir. Özellikle iskemik inme, beyin kan akışındaki tıkanıklık sonucu nörolojik hasara yol açar. Erken tanı ve tedavi, inme yönetiminde kritiktir ve Manyetik Rezonans Görüntüleme (MRG) bu aşamada önemli bir rol oynamaktadır. Ancak MRG yorumlamasının manuel olması, zaman alıcı ve uzman radyologlara bağımlı bir süreçtir. Bu durum, sağlık hizmetlerindeki yapay zeka tabanlı lezyon saptama yöntemlerine olan ihtiyacı ortaya koymaktadır. Araştırmada, inme lezyonlarının sınırlayıcı kutular ile tespiti için YOLOv8, YOLO11 ve RT-DETR gibi tespit modelleri ile DeepLabv3+ ve SegFormer gibi segmentasyon modellerinin performansını değerlendirilmektedir. Model performansını ve veri kalitesini artırmak amacıyla 3B’den 2B’ye görüntü dönüştürme, yoğunluk normalizasyonu ve lezyon maskesi iyileştirme gibi kapsamlı ön işleme teknikleri kullanılmıştır. Modeller Anatomical Tracings of Lesions After Stroke (ATLAS_2.0), ve Ischemic Stroke Lesion Segmentation (ISLES’22) olmak üzere 2 açık erişimli veri kümesi üzerinde F1-skoru ve ortalama hassasiyet (mAP) metrikleri kullanılarak karşılaştırılmıştır. Bu çalışma, her bir yaklaşımın avantajlarını ve dezavantajlarını belirleyerek, yapay zeka destekli lezyon tespitiyle klinik uygulamalar arasındaki boşluğu kapatmayı ve daha verimli ve doğru inme tanısına katkıda bulunmayı hedeflemektedir.

Anahtar Kelimeler

• İnme lezyon segmentasyonu
• beyin MRI analizi
• bilgisayarla görme
• YOLO
• derin öğrenme

COMPARATIVE ANALYSIS OF DEEP LEARNING MODELS FOR ISCHEMIC STROKE SEGMENTATION IN MRI

Öz

Stroke is one of the major global health concerns and a leading cause of death and long-term disability. Ischemic stroke occurs when a blockage in cerebral blood flow results in neurological impairment. Early diagnosis and treatment are critical in stroke management, and Magnetic Resonance Imaging (MRI) plays an important role in this phase. However, manual MRI interpretation is highly dependent on expert radiologists and time-consuming, highlighting the need for AI-based lesion detection methods in healthcare. This study evaluates the performance of various models for ischemic stroke lesion detection: YOLOv8, YOLO11, and RT-DETR for object detection, and DeepLabv3+ and SegFormer for segmentation. To enhance data quality and model performance, a comprehensive preprocessing pipeline was implemented. This included 3D to 2D image conversion, intensity normalization, and lesion mask refinement. The models are compared using F1-score and mean precision (mAP) metrics on two public datasets: Anatomical Tracings of Lesions After Stroke (ATLAS_2.0) and Ischemic Stroke Lesion Segmentation (ISLES’22). This study aims to bridge the gap between AI-driven lesion detection and real-world clinical applications by determining the advantages and disadvantages of each approach and contributing to more efficient and accurate stroke diagnosis.

Anahtar Kelimeler

• Stroke lesion segmentation
• brain MRI analysis
• computer vision
• YOLO
• deep learning

Kaynakça

1. Ahmed, R., Al Shehhi, A., Werghi, N., & Seghier, M. L. (2024). Segmentation of stroke lesions using transformers‐augmented MRI analysis. Human Brain Mapping, 45(11), e26803. https://doi.org/10.1002/hbm.26803
2. Alshehri, F., & Muhammad, G. (2023). A few-shot learning-based ischemic stroke segmentation system using weighted MRI fusion. Image and Vision Computing, 140, 104865. https://doi.org/10.1016/j.imavis.2023.104865
3. Başaran, E., Cömert, Z., Şengür, A., Budak, Ü., Çelik, Y., & Toğaçar, M. (2019). Chronic tympanic membrane diagnosis based on deep convolutional neural network. In 2019 4th international conference on computer science and engineering (UBMK) (pp. 1-4). Ieee. https://doi.org/10.1109/UBMK.2019.8907070
4. Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., & Zagoruyko, S. (2020). End-to-end object detection with transformers. In A. Vedaldi, H. Bischof, T. Brox, & J.-M. Frahm (Eds.), Computer vision – ECCV 2020, 12346, p213–229, Springer. https://doi.org/10.1007/978-3-030-58452-8_13
5. Chen, L.-C., Papandreou, G., Kokkinos, I., Murphy, K., & Yuille, A. L. (2018). DeepLab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs. IEEE Transactions on Pattern Analysis and Machine Intelligence, 40(4), 834–848. https://doi.org/10.1109/TPAMI.2017.2699184
6. Chen, S., Duan, J., Zhang, N., Qi, M., Li, J., Wang, H., Wang, R., Ju, R., Duan, Y., & Qi, S. (2023). MSA-YOLOv5: Multi-scale attention-based YOLOv5 for automatic detection of acute ischemic stroke from multi-modality MRI images. Computers in Biology and Medicine, 165, 107471. https://doi.org/10.1016/j.compbiomed.2023.107471
7. Cekic, E., Pinar, E., Pinar, M., & Dagcinar, A. (2024). Deep learning-assisted segmentation and classification of brain tumor types on magnetic resonance and surgical microscope images. World Neurosurgery, 182, e196-e204. https://doi.org/10.1016/j.wneu.2023.11.073
8. Daoudi, R., Mouelhi, A., & Sayadi, M. (2020). Automatic ischemic stroke lesions segmentation in multimodality MRI using mask region-based convolutional neural network. In 2020 4th International Conference on Advanced Systems and Emergent Technologies (IC_ASET), 362–366. IEEE. https://doi.org/10.1109/IC_ASET49463.2020.9318265

9. Elhanashi, A., Dini, P., Saponara, S., & Zheng, Q. (2024). TeleStroke: Real-time stroke detection with federated learning and YOLOv8 on edge devices. Journal of Real-Time Image Processing. 21(4), 121. https://doi.org/10.1007/s11554-024-01500-1
10. Feigin, V. L., Brainin, M., Norrving, B., Martins, S. O., Pandian, J., Lindsay, P., Grupper, M. F., & Rautalin, I. (2025). World Stroke Organization: Global stroke fact sheet 2025. International Journal of Stroke. 20(2), 132–144. https://doi.org/10.1177/17474930241308142
11. Garcia-Salgado, B. P., Almaraz-Damian, J. A., Cervantes-Chavarria, O., Ponomaryov, V., Reyes-Reyes, R., Cruz-Ramos, C., & Sadovnychiy, S. (2024). Enhanced ischemic stroke lesion segmentation in MRI using attention U-Net with generalized Dice focal loss. Applied Sciences. 14(18), 8183. https://doi.org/10.3390/app14188183
12. Geetha, R., Priya, E., & Vijayakumar, M. (2025). An approach for automated acute cerebral ischemic stroke lesion segmentation and correlation of significant features with modified Rankin scale. Biomedical Signal Processing and Control. 100, 106921. https://doi.org/10.1016/j.bspc.2024.106921
13. Gülcan, S. N., Tilki, E., Acar, K., Aktaş, A., Altınel, A. B., & Pınar, M. (2025, September). A Comparative Analysis of Convolutional Neural Networks, Vision Transformers and Hybrid Deep Learning Architectures for Lung CT Image Classification. In 2025 10th International Conference on Computer Science and Engineering (UBMK), 946-951. https://doi.org/10.1109/UBMK67458.2025.11206987
14. Hernandez Petzsche, M. R., De La Rosa, E., Hanning, U., Wiest, R., Valenzuela, W., Reyes, M., Meyer, M., Liew, S.-L., Kofler, F., Ezhov, I., Robben, D., ... Kirschke, J. S. (2022). ISLES 2022: A multi-center magnetic resonance imaging stroke lesion segmentation dataset. Scientific Data, 9, 762. https://doi.org/10.1038/s41597-022-01875-5
15. Kabakuş, A. T., & Erdoğmuş, P. (2024). Empowering self-detection: A graphical user interface powered by machine learning for early diagnosis of Alzheimer’s disease. İstanbul Ticaret Üniversitesi Fen Bilimleri Dergisi, 23(46), 245–270. https://doi.org/10.55071/ticaretfbd.1416508
16. Karthik, R., Menaka, R., Hariharan, M., & Won, D. (2021). Ischemic lesion segmentation using ensemble of multi-scale region aligned CNN. Computer Methods and Programs in Biomedicine, 200, 105831. https://doi.org/10.1016/j.cmpb.2020.105831
17. Lalinia, M., & Sahafi, A. (2024). Colorectal polyp detection in colonoscopy images using YOLO-V8 network. Signal, Image and Video Processing, 18(3), 2047–2058. https://doi.org/10.1007/s11760-023-02835-1
18. Liew, S.-L., Lo, B. P., Donnelly, M. R., Zavaliangos-Petropulu, A., Jeong, J. N., Barisano, G., Hutton, A., Simon, J. P., Juliano, J. M., Suri, A., ... Yu, C. (2022). A large, curated, open-source stroke neuroimaging dataset to improve lesion segmentation algorithms. Scientific Data, 9, 320. https://doi.org/10.1038/s41597-022-01401-7
19. Müller, D., Soto-Rey, I., & Kramer, F. (2022). Towards a guideline for evaluation metrics in medical image segmentation. BMC Research Notes, 15(1), 210. https://doi.org/10.1186/s13104-022-06096-y
20. Nazari-Farsani, S., Nyman, M., Karjalainen, T., Bucci, M., Isojärvi, J., & Nummenmaa, L. (2020). Automated segmentation of acute stroke lesions using a data-driven anomaly detection on diffusion weighted MRI. Journal of Neuroscience Methods, 333, 108575. https://doi.org/10.1016/j.jneumeth.2019.108575
21. Nouman, M., Mabrok, M., & Rashed, E. A. (2024). Neuro-TransUNet: A comprehensive transformer-based architecture for precise segmentation of stroke lesions in 3D MRI. In Proceedings of the 2024 9th International Conference on Multimedia and Image Processing, 152–156. ACM. https://doi.org/10.1145/3665026.3665049
22. Pinar, M., Altinel, A. B., & Aktas, A. (2025). HeaRT: An Innovative Health Representation Transformer With Clustered Feature Augmentation for Biomedical Text Classification. IEEE Access, 13, 215748-215770. https://doi.org/10.1109/access.2025.3646756
23. Rahman, A., Chowdhury, M. E. H., Ibne Wadud, M. S., Sarmun, R., Mushtak, A., Zoghoul, S. B., & Al-Hashimi, I. (2025). Deep learning-driven segmentation of ischemic stroke lesions using multi-channel MRI. Biomedical Signal Processing and Control, 105, 107676. https://doi.org/10.1016/j.bspc.2025.107676
24. Rani, S., Singh, B. K., Koundal, D., & Athavale, V. A. (2022). Localization of stroke lesion in MRI images using object detection techniques: A comprehensive review. Neuroscience Informatics, 2(3), 100070. https://doi.org/10.1016/j.neuri.2022.100070
25. Ragab, M. G., Danyaro, K. U., Abdulkadir, S. J., Al-Selwi, S. M., Sumiea, E. H., & Alqushaibi, A. (2024). Brain tumor detection instance segmentation using YOLOv8 in medical imaging. In 2024 IEEE Asia-Pacific Conference on Applied Electromagnetics(APACE), 55–58. https://doi.org/10.1109/APACE62360.2024.10877358
26. Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 779–788. IEEE. https://doi.org/10.1109/CVPR.2016.91
27. Ren, S., He, K., Girshick, R., & Sun, J. (2017). Faster R-CNN: Towards real-time object detection with region proposal networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6), 1137–1149. https://doi.org/10.1109/TPAMI.2016.2577031
28. Saito, T., & Rehmsmeier, M. (2015). The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLOS ONE, 10(3), e0118432. https://doi.org/10.1371/journal.pone.0118432
29. Selen, İ., Erol, Ü., Aktaş, A., Altınel, A. B., & Pınar, M. (2025, September). AI-Assisted Detection of Dental Findings in Panoramic Dental Radiographs. In 2025 10th International Conference on Computer Science and Engineering (UBMK) (pp. 883-888). IEEE. https://doi.org/10.1109/UBMK67458.2025.11206840
30. Tekin, B., & Gurbanov, R. (2023). ALPHAFOLD: Revolutionizing protein folding through deep learning and neural networks. İstanbul Ticaret Üniversitesi Fen Bilimleri Dergisi, 22(44), 445–466. https://doi.org/10.55071/ticaretfbd.1323165
31. Toğaçar, M., Cömert, Z., & Ergen, B. (2021). Enhancing of dataset using DeepDream, fuzzy color image enhancement and hypercolumn techniques to detection of the Alzheimer's disease stages by deep learning model. Neural Computing and Applications, 33(16), 9877-9889. https://doi.org/10.1007/s00521-021-05758-5
32. Ultralytics. (2023). YOLOv8. https://docs.ultralytics.com/models/yolov8
33. Ultralytics. (2024). YOLO11. https://docs.ultralytics.com/tr/models/yolo11
34. Wang, G., Chen, Y., An, P., Hong, H., Hu, J., & Huang, T. (2023). UAV-YOLOv8: A small-object-detection model based on improved YOLOv8 for UAV aerial photography scenarios. Sensors, 23(16), 7190. https://doi.org/10.3390/s23167190
35. Wu, Z., Zhang, X., Li, F., Wang, S., Huang, L., & Li, J. (2023). W-Net: A boundary-enhanced segmentation network for stroke lesions. Expert Systems with Applications, 230, 120637. https://doi.org/10.1016/j.eswa.2023.120637
36. Xie, E., Wang, W., Yu, Z., Anandkumar, A., Alvarez, J. M., & Luo, P. (2021). SegFormer: Simple and efficient design for semantic segmentation with transformers. Advances in Neural Information Processing Systems, 34, 12077–12090.
37. Xue, Y., Farhat, F. G., Boukrina, O., Barrett, A. M., Binder, J. R., Roshan, U. W., & Graves, W. W. (2020). A multi-path 2.5 dimensional convolutional neural network system for segmenting stroke lesions in brain MRI images. NeuroImage: Clinical, 25, 102118. https://doi.org/10.1016/j.nicl.2019.102118
38. Yu, Y., Heit, J. J., & Zaharchuk, G. (2021). Improving ischemic stroke care with MRI and deep learning artificial intelligence. Topics in Magnetic Resonance Imaging, 30(4), 187–195. https://doi.org/10.1097/RMR.0000000000000290
39. Zhang, S., Xu, S., Tan, L., Wang, H., & Meng, J. (2021). Stroke lesion detection and analysis in MRI images based on deep learning. Journal of Healthcare Engineering, 5524769. https://doi.org/10.1155/2021/5524769
40. Zhang, S., Tan, L., Han, Q., Wang, H., & Meng, J. (2023). Automatic report generation on a large-scale stroke MRI dataset. In 2023 IEEE 6th International Conference on Electronic Information and Communication Technology (ICEICT) (pp. 123–128). IEEE. https://doi.org/10.1109/ICEICT57916.2023.10245487
41. Zhao, Y., Lv, W., Xu, S., Wei, J., Wang, G., Dang, Q., Liu, Y., & Chen, J. (2024). DETRs beat YOLOs on real-time object detection. In 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 16965–16974). IEEE. https://doi.org/10.1109/CVPR52733.2024.01605
42. Zhou, Z. H. (2025). Ensemble methods: foundations and algorithms. CRC press.