Retina Damar Segmentasyonunda Sinir Mimarisi Arama için Farklı Kodlama Yöntemlerinin Karşılaştırılması

Yıl 2023, Cilt: 35 Sayı: 4, 447 - 459, 31.12.2023

Zeki Kuş , Berna Kiraz

https://doi.org/10.7240/jeps.1335157

Öz

Retinal kan damar segmentasyonu diyabetik retinopati ve yaşa bağlı makula dejenerasyonu gibi göz hastalıklarının tespiti ve incelemesi açısından kritik bir görevdir. U-şekilli derin sinir ağlarının bu görev için başarılı sonuçlar verdiği bilinmektedir; fakat bu ağların optimize edilmesi gereken bir çok hiper-parametresi bulunmaktadır. Bu ağların otomatik bir şekilde optimizasyonu ve aranması için birçok sinir mimarisi arama (SMA) çalışması gerçekleştirilmiştir. SMA çalışmaları incelendiğinde seçilen kodlama şemalarının üretilen ağların karmaşıklığını ve performansını doğrudan etkilediği görülmüştür. Bu çalışmada, retinal kan damar segmentasyonu için sunduğumuz iki SMA çalışmasında (UNAS-Net ve MedUNAS) önerilen kodlama şemalarının performansları herkese açık olarak yayınlanan iki farklı retinal kan damar segmentasyonu veri kümesi üzerinde karşılaştırılmıştır. Elde edilen sonuçlara bakıldığında, önerilen her iki yöntemin temel alınan U-Net'ten 25 kata kadar daha az parametre ile tüm ölçütler açısından daha iyi performans gösterdiği görülmüştür. Ayrıca, UNAS-Net ve MedUNAS'ın SMA çalışmaları arasında en az parametre ile yüksek rekabetçi sonuçlar elde edebildiği gösterilmiştir.

Anahtar Kelimeler

Retina Damar Segmentasyonu, Sinir Mimarisi Arama, Kodlama, Karşitlik Temelli Diferansiyel Gelişim

Kaynakça

• [1] Abràmoff, M. D., Garvin, M. K., & Sonka, M. (2010). Retinal imaging and image analysis. IEEE reviews in biomedical engineering, 3, 169-208.
• [2] Wang, B., Qiu, S., He, H. (2019). Dual Encoding U-Net for Retinal Vessel Segmentation. In: Shen, D., et al. Medical Image Computing and Computer Assisted Intervention – MICCAI 2019. MICCAI 2019. Lecture Notes in Computer Science(), vol 11764. Springer, Cham.
• [3] Ronneberger, O., Fischer, P., Brox, T. (2015). U-Net: Convolutional Networks for Biomedical Image Segmentation. In: Navab, N., Hornegger, J., Wells, W., Frangi, A. (eds) Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015. MICCAI 2015. Lecture Notes in Computer Science(), vol 9351. Springer, Cham.
• [4] Chen, C., Chuah, J. H., Ali, R., & Wang, Y. (2021). Retinal vessel segmentation using deep learning: a review. IEEE Access, 9, 111985-112004.
• [5] Zhou, Y., Yu, H., & Shi, H. (2021). Study group learning: Improving retinal vessel segmentation trained with noisy labels. In Medical Image Computing and Computer Assisted Intervention–MICCAI 2021: 24th International Conference, Strasbourg, France, September 27–October 1, 2021, Proceedings, Part I 24 (pp. 57-67). Springer International Publishing.
• [6] Liu, W., Yang, H., Tian, T., Cao, Z., Pan, X., Xu, W., ... & Gao, F. (2022). Full-resolution network and dual-threshold iteration for retinal vessel and coronary angiograph segmentation. IEEE Journal of Biomedical and Health Informatics, 26(9), 4623-4634.
• [7] Rong, Y., Xiong, Y., Li, C., Chen, Y., Wei, P., Wei, C., & Fan, Z. (2023). Segmentation of retinal vessels in fundus images based on U-Net with self-calibrated convolutions and spatial attention modules. Medical & Biological Engineering & Computing, 1-11.
• [8] Baker, B., Gupta, O., Naik, N., & Raskar, R. (2016). Designing neural network architectures using reinforcement learning. arXiv preprint arXiv:1611.02167.
• [9] Zoph, B., & Le, Q. V. (2016). Neural architecture search with reinforcement learning. arXiv preprint arXiv:1611.01578.
• [10] Grathwohl, W., Creager, E., Ghasemipour, S. K. S., & Zemel, R. (2018). Gradient-based optimization of neural network architecture.
• [11] Liu, H., Simonyan, K., & Yang, Y. (2018). Darts: Differentiable architecture search. arXiv preprint arXiv:1806.09055.
• [12] Liu, X., Song, L., Liu, S., & Zhang, Y. (2021). A review of deep-learning-based medical image segmentation methods. Sustainability, 13(3), 1224.
• [13] Houreh, Y., Mahdinejad, M., Naredo, E., Dias, D. M., & Ryan, C. (2021). HNAS: Hyper Neural Architecture Search for Image Segmentation. In ICAART (2) (pp. 246-256).
• [14] Wei, J., Zhu, G., Fan, Z., Liu, J., Rong, Y., Mo, J., ... & Chen, X. (2021). Genetic U-Net: automatically designed deep networks for retinal vessel segmentation using a genetic algorithm. IEEE Transactions on Medical Imaging, 41(2), 292-307.
• [15] Rajesh, C., Sadam, R., & Kumar, S. (2023). An evolutionary U-shaped network for Retinal Vessel Segmentation using Binary Teaching–Learning-Based Optimization. Biomedical Signal Processing and Control, 83, 104669.
• [16] Kuş, Z., Aydın, M., Kiraz, B., & Can, B. (2022). Neural Architecture Search Using Metaheuristics for Automated Cell Segmentation. In Metaheuristics International Conference (pp. 158-171). Cham: Springer International Publishing.
• [17] Kuş, Z., Kiraz, B. (2023). Evolutionary Architecture Optimization for Retinal Vessel Segmentation. IEEE Journal of Biomedical and Health Informatics.
• [18] Dong, J., Hou, B., Feng, L., Tang, H., Tan, K. C., & Ong, Y. S. (2022). A cell-based fast memetic algorithm for automated convolutional neural architecture design. IEEE Transactions on Neural Networks and Learning Systems.
• [19] Awad, N., Mallik, N., & Hutter, F. (2020). Differential evolution for neural architecture search. arXiv preprint arXiv:2012.06400.
• [20] Rahnamayan, S., Tizhoosh, H. R., & Salama, M. M. (2008). Opposition-based differential evolution. IEEE Transactions on Evolutionary computation, 12(1), 64-79.
• [21] Staal, J., Abràmoff, M. D., Niemeijer, M., Viergever, M. A., & Van Ginneken, B. (2004). Ridge-based vessel segmentation in color images of the retina. IEEE transactions on medical imaging, 23(4), 501-509.
• [22] Carballal, A., Novoa, F. J., Fernandez-Lozano, C., García-Guimaraes, M., Aldama-López, G., Calviño-Santos, R., ... & Pazos, A. (2018). Automatic multiscale vascular image segmentation algorithm for coronary angiography. Biomedical Signal Processing and Control, 46, 1-9.
• [23] Li, J., Gao, G., Liu, Y., & Yang, L. (2023). MAGF-Net: A multiscale attention-guided fusion network for retinal vessel segmentation. Measurement, 206, 112316.
• [24] Xie, X., Song, X., Lv, Z., Yen, G. G., Ding, W., & Sun, Y. (2023). Efficient evaluation methods for neural architecture search: A survey. arXiv preprint arXiv:2301.05919.
• [25] Kuş, Z., Kiraz, B., Göksu, T. K., Aydın, M., Özkan, E., Vural, A., Kiraz, A., & Can, B. (2023). Differential evolution-based neural architecture search for brain vessel segmentation. In Engineering Science and Technology, an International Journal (Vol. 46, p. 101502).
• [26] Popat, V., Mahdinejad, M., Cedeño, O. D., Naredo, E., & Ryan, C. (2020). GA-based U-Net Architecture Optimization Applied to Retina Blood Vessel Segmentation. In IJCCI (pp. 192-199).
• [27] Zhou, Z., Rahman Siddiquee, M. M., Tajbakhsh, N., & Liang, J. (2018). Unet++: A nested u-net architecture for medical image segmentation. In Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support: 4th International Workshop, DLMIA 2018, and 8th International Workshop, ML-CDS 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 20, 2018, Proceedings 4 (pp. 3-11). Springer International Publishing.
• [28] Oktay, O., Schlemper, J., Folgoc, L. L., Lee, M., Heinrich, M., Misawa, K., ... & Rueckert, D. (1804). Attention u-net: Learning where to look for the pancreas. arXiv 2018. arXiv preprint arXiv:1804.03999.
• [29] Mou, L., Zhao, Y., Chen, L., Cheng, J., Gu, Z., Hao, H., ... & Liu, J. (2019). CS-Net: Channel and spatial attention network for curvilinear structure segmentation. In Medical Image Computing and Computer Assisted Intervention–MICCAI 2019: 22nd International Conference, Shenzhen, China, October 13–17, 2019, Proceedings, Part I 22 (pp. 721-730). Springer International Publishing.
• [30] Zhang, S., Fu, H., Yan, Y., Zhang, Y., Wu, Q., Yang, M., ... & Xu, Y. (2019). Attention guided network for retinal image segmentation. In Medical Image Computing and Computer Assisted Intervention–MICCAI 2019: 22nd International Conference, Shenzhen, China, October 13–17, 2019, Proceedings, Part I 22 (pp. 797-805). Springer International Publishing.
• [31] Wang, W., Zhong, J., Wu, H., Wen, Z., & Qin, J. (2020). Rvseg-net: An efficient feature pyramid cascade network for retinal vessel segmentation. In Medical Image Computing and Computer Assisted Intervention–MICCAI 2020: 23rd International Conference, Lima, Peru, October 4–8, 2020, Proceedings, Part V 23 (pp. 796-805). Springer International Publishing.
• [32] Wu, H., Wang, W., Zhong, J., Lei, B., Wen, Z., & Qin, J. (2021). Scs-net: A scale and context sensitive network for retinal vessel segmentation. Medical Image Analysis, 70, 102025.