Kan Damarı ve Optik Disk Bölütlemesi için Karar Destek Sistemi

Abstract

Son zamanlarda şeker hastalığı hızla artıyor. Şeker hastalığının farklı organları etkileyen birçok türü vardır. Diyabetik retinopi, diyabet türlerinden biridir. Diyabetik retinopi ve glokom, körlüğe yol açabilen göz hastalıklarıdır. Optik disk segmentasyonu bu hastalıkların belirlenmesinde faydalıdır. Ancak burada aydınlatma varyasyonları gibi bazı engeller ortaya çıkmaktadır. Fundus görüntülerinde kan damarlarının etkisi giderilmelidir. Bunun nedeni, kan damarlarının optik diskin (OD) kenarlarını kesebilmesidir. Bu çalışmada amaç, kan damarlarını görüntülerden çıkarmak ve ardından optik diski başarılı bir şekilde bölütlemektir. LinkNet'e dayalı LinkNetRCB7 adlı yeni bir yaklaşım geliştirdik. LinkNetRCB7 başarılı sonuçlar elde etti. Kan damarları segmentasyonu ve OD doğruluğu, STARE veri setinde %98,5 ve DRISHTI GS'de %98.850 olarak hesaplandı. Bu aşamaları içeren bir karar destek sistemi çalışma kapsamında önerilmiştir. Hastalık teşhisinde karar destek sistemlerinin kullanımı yaygınlaşmaktadır. Literatürde diyabetik retinopi için derin öğrenme ve görüntü işleme algoritmalarını içeren bir karar destek sistemi görülmemektedir. Önerilen karar destek sistemi ile diyabetik retinopi’nin teşhis sürecinde karar vericilere yardımcı olabilecek bir sistem tasarlanmıştır.

Keywords

• LinkNet
• Karar Destek Sistemi
• Derin Öğrenme
• ResNetC
• Diyabet
• Diyabetik Retinopi

Decision Support System for Blood Vessel and Optic Disc Segmentation

Abstract

Recently, diabetes is rapidly increasing. Diabetes has many types that infect different organs. Diabetic retinotopic is one of them. Diabetic retinotopic and glaucoma are eye diseases that can lead to blindness. Optic disc segmentation helps identify these diseases. However, here some obstacles arise, such as illumination variations. The effect of blood vessels in fundus images should be removed. This is because blood vessels can incise the edges of the optic disc (O.D.). This study aimed to remove blood vessels from images and then successfully segment the optic disc. We developed a new approach called LinkNetRCB7 based on LinkNet. LinkNetRCB7 has achieved successful results. The accuracy for blood vessel segmentation and O.D. was calculated to be 98.5% on the Stare dataset and 98.850% on DRISHTI GS. A decision support system (DSS) including these stages has been proposed within the scope of the study. The use of DSS in disease diagnosing is becoming widespread. No decision support system in the literature includes deep learning and image processing algorithms for diabetic retinopia. With the proposed decision support system, a system that can help decision-makers diagnose diabetic retinotopic has been designed.

Keywords

• LinkNet
• Decision Support System
• Deep Learning
• ResNetC
• Diabetes
• Diabetic Retinopia

References

1. [1] K. P. Mali, B. T. Jadhav, and I. K. Mujawar. “Study of Diabetic retinopia Detection Using Deep Learning Techniques,” International Research Journal of Humanities and Interdisciplinary Studies Journal, Special Issue, pp. 208-216, March 2022. Doi:03.2021-11278686.
2. [2] Y. Guo, and Y. Peng. ”CARNet: Cascade attentive RefineNet for multi-lesion segmentation of diabetic retinopia images,” Complex Intelligent Systems, vol.8, pp. 1681–1701, January 2022. Doi: 10.1007/s40747-021-00630-4.
3. [3] G. Zhang, B. Sun, Z. Chen, Y. Gao, Z. Zhang, W. Yang, and K. Li. ”Diabetic retinopia Grading by Deep Graph Correlation Network on Retinal Images Without Manual Annotations,” Frontiers in Medicine, vol.9, pp. 1-9, April 2022. Doi: 10.3389/fmed.2022.872214.
4. [4] S. E. Mansour, D. J. Browning, K. Wong, H. W. Flynn, and A. R. Bhavsar. ”The Evolving Treatment of Diabetic retinopia,” Clin Ophthalmol, vol.14, pp. 653–678, March 2020. Doi: 10.2147/OPTH.S236637.
5. [5] S. H. Sinclair, and S. S. Schwartz. ”Diabetic retinopia–An Underdiagnosed and Undertreated Inflammatory, Neuro Vascular Complication of Diabetes,” Frontiers in Medicine, vol.10, December 2019. Doi: 10.3389/fendo.2019.00843.
6. [6] S. Tang, Z. Qi, J. Granley, and M. Beyeler. ”U-Net with Hierarchical Bottleneck Attention for Landmark Detection in Fundus Images of the Degenerated Retina,” in Proc. International Workshop on Ophthalmic Medical Image Analysis,” pp. 62-71, Jully 2021. Doi: https://doi.org/10.48550/arXiv.2107.04721.
7. [7] A. G. Marrugo, M. S. Millan. ”Optic disc segmentation in retinal images,” [Online]. Avaliable: https://iopscience.iop.org/article/10.1088/1742-6596/274/1/012039/pdf. [Accessed: Sept, 7, 2022]
8. [8] C. Akyel and N. Arıcı. “A New Approach to Hair Noise Cleaning and Lesion Segmentation in Images of Skin Cancer,” Journal of Polytechnic, vol. 23, no. 3, pp. 821-828, April 2020. Doi: https://doi.org/10.17671/gazibtd.1060330.

9. [9] S. Kumar, G. Sathyadevi, and S.Sivanesh. ”Decision Support System for Medical Diagnosis Using Data Mining,”[Online]. Avaliable: https://www.researchgate.net/publication/267934132_Decision_Support_System_for_Medical_Diagnosis_Using_Data_Mining. [Accessed: Sept, 7, 2022]
10. [10] M. J. Aqel, O. A. Nakshabandi, A. Adeniyi. ”Decision Support Systems Classification in Industry,”. [Online]. Avaliable: http://pen.ius.edu.ba/index.php/pen/article/view/550/343 [Accessed: Sept, 7, 2022]
11. [11] M. Maison, T Lestari, and A. Luthfi. ”Retinal Blood Vessel Segmentation using Gaussian Filter,” [Online]. Avaliable: https://iopscience.iop.org/article/10.1088/1742-6596/1376/1/012023 [Accessed: Sept, 7, 2022]
12. [12] Y. Yang, F. Shao, Z. Fu, and R. Fu. ”Blood vessel segmentation of fundus images via cross-modality dictionary learning,” Applied Optics, vol. 57, no. 25, September 2018. Doi: https://doi.org/10.1364/AO.57.007287.
13. [13] Y. Yang, F. Shao, Z. Fu, and R. Fu. ”Discriminative dictionary learning for retinal vessel segmentation using fusion of multiple features,” Signal, Image and Video Processing, vol.13, pp. 1529–1537, 2019.
14. [14] Y. Jiang, C. Wu, G. Wang, H. Yao, and W. Liu. ”MFI-Net: A multi-resolution fusion input network for retinal vessel segmentation,” PLOS ONE, vol. 16, no. 7, May 2021. Doi: 10.1007/s11760-019-01501-9.
15. [15] W. Zhou, Y. Yi, Y. Gao, and J. Dai. ”Optic Disc and Cup Segmentation in Retinal Images for Glaucoma Diagnosis by Locally Statistical Active Contour Model with Structure Prior,” Hindawi Computational and Mathematical Methods in Medicine, vol. 2019, 2019. Doi: 10.1155/2019/8973287.
16. [16] Y. Ma, X. Li, X. Duan, Y. Peng, and Y. Zhang. ”Retinal Vessel Segmentation by Deep Residual Learning with Wide Activation,” Hindawi Computational Intelligence and Neuroscience, vol. 2020, 2020. Doi: 10.1155/2020/8822407.
17. [17] M. Li, Q. Yin, and M. Lu, ”Retinal Blood Vessel Segmentation Based on Multi-Scale Deep Learning,” [Online]. Avaliable: https://ieeexplore.ieee.org/document/8511244 [Accessed: Sept, 7, 2022].
18. [18] H. Boudegga, Y. Elloumi, M. Akil, M. H. Bedoui, R. Kachouri, A. B. Abdallah, ”Fast and efficient retinal blood vessel segmentation method based on deep learning network,” Computerized Medical Imaging and Graphics, vol. 90, June 2021. Doi: 10.1016/j.compmedimag.2021.101902.
19. [19] L. S. Brea, D. A. De Jesus, S. Klein, and T. V. Walsum, ”Deep learning-based retinal vessel segmentation with cross-modal evaluation,” [Online]. Avaliable: http://proceedings.mlr.press/v121/sanchez-brea20a/sanchez-brea20a.pdf. [Accessed: Sept, 7, 2022].
20. [20] B. Browatzki, J. Lies, and C. Wallraven, ”Encoder-Decoder Networks for Retinal Vessel Segmentation Using Large Multi-scale Patches,” [Online]. Avaliable: https://dl.acm.org/doi/abs/10.1007/978-3-030-63419-3_5. [Accessed: Sept, 7, 2022].
21. [21] S. Guo, K. Wang, H. Kang, Y. Zhang, Y. Gao, T. Li, ”BTS-DSN: Deeply Supervised Neural Network with Short Connections for Retinal Vessel Segmentation,” International Journal of Medical Informatics, vol. 126, pp. 105-113, June 2019. Doi: 10.1016/j.ijmedinf.2019.03.015.
22. [22] Z. Li, M. Jia, X. Yang, and M. Xu, ”Blood Vessel Segmentation of Retinal Image Based on Dense-U-Net Network,” Micromachines, vol. 12, no. 12, December 2021. Doi: 10.3390/mi12121478
23. [23] A. Almazroa, W. Sun, S. Alodhayb, K. Raahemifar, and V. Lakshminarayanan, ”Optic disc segmentation for glaucoma screening system using fundus images,” Clinical Ophthalmology, vol. 2017, pp. 2017-2029, November 2017. Doi: 10.2147/OPTH.S140061.
24. [24] T. Laibacher, T. Weyde, S. Jalali. ”M2U-Net: Effective and Efficient Retinal Vessel Segmentation for Real-World Applications,” in Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 16-17 June 2019, Long Beach, CA, USA [Online]. Avaliable: https://arxiv.org/abs/1811.07738. [Accessed: Sept, 7, 2022].
25. [25] W. Liu, H. Lei, H. Xie, B. Zhao, G. Yue, and B. Lei, ”Multi-level Light U-Net and Atrous Spatial Pyramid Pooling for Optic Disc Segmentation on Fundus Image,” in proc. Ophthalmic Medical Image Analysis 7th International Workshop, pp. 104–113, 8 oct 2020, Lima, Peru [Online]. Avaliable: https://link.springer.com/chapter/10.1007/978-3-030-63419-3_11. [Accessed: Sept, 7, 2022].
26. [26] H.N. Veena, A. Muruganandham, T. Kumaran, ”A novel optic disc and optic cup segmentation technique to diagnose glaucoma using deep learning convolutional neural network over retinal fundus images,” Journal of King Saud University – Computer and Information Science, [Online]. Avaliable: https://www.sciencedirect.com/science/article/pii/S1319157821000422?via%3Dihub. [Accessed: Sept, 7, 2022].
27. [27] M. Tabassum, T. Khan, M. Arsalan, S. Naqvi, M. Ahmed, H. Madni, And J. Mirza, ”Cded-Net: Joint Segmentation of Optic Disc and Optic Cup for Glaucoma Screening,” IEEE, vol.8 , pp. 102733 - 102747, May 2020. Doi: 10.1109/ACCESS.2020.2998635
28. [28] H. He, L. Lin, X. Tang and Z. Cai, ”JOINED : Prior Guided Multi-task Learning for Joint Optic Disc/Cup Segmentation and Fovea Detection,” [Online]. Avaliable: https://arxiv.org/abs/2203.00461 [Accessed: Sept, 7, 2022].
29. [29] Z. Erdoğan, ”U-Net Architecture Optimization For Optic Disc Segmentation In Retinal Images,” Master Thesis, Department of Computer Engineering Computer Engineering Programme, Istanbul Technical University, İstanbul, Turkey.
30. [30] B. Liu, D. Pan, and H. Song, ”Joint optic disc and cup segmentation based on densely connected depthwise separable convolution deep network,” BMC Medical Imaging, vol.21, no.14, January 2021. Doi: 10.1186/s12880-020-00528-6.
31. [31] Z. Tian, Y. Zheng, X. Li, S. Du, And X. Xu, ”Graph convolutional network based optic disc and cup segmentation on fundus images,” Biomedical Optics Express, vol. 11, no. 6, pp. 3043-3057, June 2020. Doi: 10.1364/BOE.390056.
32. [32] R. Kamble, P. Samanta, and N. Singhal, Optic Disc, ”Cup and Fovea Detection from Retinal Images Using UNet++ with EfficientNet Encoder,” In Proc. of the International Workshop on Ophthalmic Medical Image Analysis, 8 October 2020, Lima, Peru. [Online]. Avaliable: https://link.springer.com/chapter/10.1007/978-3-030-63419-3_10 [Accessed: Sept, 7, 2022].
33. [33] J. Lin, X. Liao, L. Yu, and J. Pan, ”Res-UNet Based Optic Disk Segmentation in Retinal Image,” Journal of Computers, vol. 31 no. 3, pp. 183-194, March 2020. Doi: 10.3966/199115992020063103014
34. [34] D. Božić-Štulić, M. Braović, and M. Braović, ”Deep learning based approach for optic disc and optic cup semantic segmentation for glaucoma analysis in retinal fundus images,” International Journal of Electrical and Computer Engineering Systems, vol 11, no 2, Jully 2020. Doi: 10.32985/ijeces.11.2.6.
35. [35] J. Orlando, H. F.U., J. Breda, K. Keer, D. Bathula, A. Pint, R. Fang, P. Hengh, J. Kim, J. Lee, J. Lee, X. Li, P. Liu, S. Lu, B. Murugesan, V. Naranjo, S. Phaye, S. Shankaranarayana, and H. Bogunovica, Refuge challenge: a unified framework for evaluating automated methods for glaucoma assessment from fundus photographs. Med Image Anal. vol. 2020, no. 59. Doi:10.1016/j.media.2019.101570.
36. [36] H. Ucuzal, ”Yapay Zekaya Dayalı Anlamsal Video İşleme Yöntemlerinin Tıpta Kullanılabilirliğinin Araştırılması,” Master Yüksek Lisans Tezi, İnönü Üniversitesi, Sağlık Bilimleri Enstitüsü, Biyoistatistik ve Tıp Bilişimi Anabilim Dalı, Malatya, Turkey. 2020.
37. [37] J. Todd, A. Gepp, ”Improving decision making in the management of hospital readmissions using modern survival analysis techniques,” Decision Support Systems, vol. 156, no.2, May 2022. Doi: 10.1016/j.dss.2022.113747
38. [38] S. Bozkurt, ”Gerçek Zamanlı Mamografi Yorumu Karar Destek Sistemi,” Biyoistatistik ve Tıbbi Bilişim Anabilim Dalı, Akdeniz Üniversitesi, Sağlık Bilimleri Enstitüsü, Doktora Tezi, Antalya, Turkey.2015.
39. [39] I. Çakmak, ”Makine Öğrenmesi Yöntemleriyle Tümör Kontrol Olasılığının Hesaplanması, Biyoistatistik ve Tıbbi Bilişim Anabilim Dalı, Sağlık Bilimleri Enstitüsü, KTÜ, Trabzon, Yüksek Lisans Tezi, Türkiye, 2017.
40. [40] M. E. Çorapçıoğlu, ”Tıpçıt: Tıbbi Karar Destek Sistemi Çekirdeği,” Bilgisayar Mühendisliği Anabilim Dalı, Yüksek Lisans Tezi, Başkent Üniversitesi, Ankara, 2006, Türkiye.
41. [41] E. Koç, ”Yöntem ve Uygulama Açısından Klinik Karar Destek Sistemleri, Fen Bilimleri Enstitüsü, Bilgisayar Mühendisliği, Okan Üniversitesi, Yüksek Lisans Tezi, İstanbul, Turkey,2013.
42. [42] Y. Park, ”Concise Logarithmic Loss Function for Robust Training of Anomaly Detection Model,” [Online]. Avaliable: https://arxiv.org/abs/2201.05748. [Accessed: Sept, 7, 2022].
43. [43] N. Savioli, G. Montana, P. Lamata, ”V-FCNN: Volumetric Fully Convolution Neural Network For Automatic Atrial Segmentation,” in Proc. International Workshop on Statistical Atlases and Computational Models of the Heart, pp. 273-281, 14 Feb. 2019.Granada, Spain. [Online]. Avaliable: https://arxiv.org/pdf/1808.01944.pdf. [Accessed: Sept, 7, 2022].
44. [44] I. Loshchilov, and F. Hutter, ”DECOUPLED WEIGHT DECAY REGULARIZATION,” in Proc. ICLR 2019 Conference homepage, pp. 273-281, 6-9 May. 2019. New Orleans, Louisiana, United States. [Online]. Avaliable: https://arxiv.org/abs/1711.05101. [Accessed: Sept, 7, 2022].
45. [45] L. Zhou, C. Zhang, M. Wu. “D-LinkNet with Pretrained The encoder and Dilated Convolution for Resolution Satellite Imagery Road Extraction,”. In Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 18–22 June 2018, Salt Lake City, UT, USA. [Online]. Avaliable: https://ieeexplore.ieee.org/document/8575492. [Accessed: Sept, 7, 2022].
46. [46] A. Chaurasia, E. Culurciello. “LinkNet: Exploiting The encoder Representations for Efficient Semantic Segmentation,”. In Proceedings of the IEEE Visual Communications and Image Processing (VCIP), 10–13 December 2017, St. Petersburg, FL, USA. [Online]. Avaliable: https://arxiv.org/abs/1707.03718. [Accessed: Sept, 7, 2022].
47. [47] S. Kallam, M. S. Kumar, V. A. Natarajan, R. Patan. “Segmentation of Nuclei in Histopathology images using Fully Convolutional Deep Neural Architecture,”. In Proceedings of the 2020 International Conference on Computing and Information Technology (ICCIT-1441), 23 November 2020, Tabuk, Saudi Arabia. [Online]. Avaliable: https://ieeexplore.ieee.org/document/9213817. [Accessed: Sept, 7, 2022].
48. [48] M. Sandler, A. G. Howard, M. Zhu, A. Zhmoginov, L. Chen. “Mobilenetv2: Inverted residuals and linear bottlenecks,”. In Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 18–23 June 2018, Salt Lake City, UT, USA. [Online]. Avaliable: https://arxiv.org/abs/1801.04381. [Accessed: Sept, 7, 2022].
49. [49] T. Mingxing, V. L. Quoc. “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,”. In Proceedings of the 36th International Conference on Machine Learning, 9–15 June 2019, Long Beach, CA, USA. [Online]. Avaliable: https://arxiv.org/abs/1905.11946. [Accessed: Sept, 7, 2022].
50. [50] M. Chetoui, M. A. Akhloufi. “Explainable Diabetic retinopia using EfficientNET,”. In Proceedings of the 42nd Annual International Conference of the IEEE Engineering in Medicine Biology Society (EMBC), 20–24 July 2020, Montreal, QC, Canada. [Online]. Avaliable: https://pubmed.ncbi.nlm.nih.gov/33018388/ [Accessed: Sept, 7, 2022].
51. [51] C. Akyel, N. Arıcı, ”Hair Removal and Lesion Segmentation with FCN8- ResNetC and Image Processing in Images of Skin Cancer,” Journal of Information Technologies, vol. 15, no. 2, pp. 231-238, 2022. Doi: 10.17671/gazibtd.1060330.
52. [52] C. Akyel, N. Arıcı, ”LinkNet-B7: Noise Removal and Lesion Segmentation in Images of Skin Cancer,” Mathematics, vol. 10, no. 5, pp. 736-751. Doi: 10.3390/math10050736.
53. [53] L. Nieradzik, G. Scheuermann, B. Saur, C. Gillmann, “Effect of the output activation function on the probabilities and errors in medical image segmentation,”. [Online]. Avaliable: https://arxiv.org/pdf/2109.00903.pdf [Accessed: Sept, 7, 2022].
54. [54] W. Xiong, X. Jia, D. Yang, M. Ai, L. Li, S. Wang. “DP-LinkNet: A convolutional network for historical document image binarization,”. Internet Inf. Syst. vol. 15, no. 5, pp. 1778–1797, May 2021. Doi: 10.3837/tiis.2021.05.011.
55. [55] M. Dönerçark, V. Tecim, ”Kurumsal Karar Destek Sistemlerinde Yapay Zeka Kullanımı: Tasarım ve Uygulama,” [Online]. Avaliable: https://dergipark.org.tr/tr/pub/ybs/issue/58550/821708#article_cite [Accessed: Sept, 7, 2022].
56. [56] S. A. Tuncer, ”Optic Disc Segmentation based on Template Matching and Active Contour Method,” Balkan Journal Of Electrical Computer Engineering, Vol. 7, No. 1, January 2019. Doi: 10.17694/bajece.470796.