Oküler Boyama Görüntülerinde Kornea Ülserinin Teşhisi İçin Derin Öğrenmeye Dayalı Bir Sınıflandırma Çalışması

Abstract

Kornea ülseri dünya genelinde yaygın görülen bir hastalık olup kornea körlüğünün önce gelen nedenlerindendir. Hastalığın teşhisi uzmanlık gerektirmekte olup, özellikle az gelişmiş ülkelerde tecrübeli oftalmolog sayısı yeterli sayıda değildir. Bu durum hastalığın teşhisinde etkin ve uzmanlara destek sistemlerin oluşturulmasını gerekli kılmaktadır. Ancak henüz bu konuda yapılmış olan çalışmaların sayısı yeterli düzeyde değildir. Bu çalışmada 712 adet ve 3 türden oluşan, oküler boyama tekniği ile elde edilen kornea ülser görüntüsü kullanılarak CNN tabanlı sınıflandırmalar gerçekleştirilmiştir. Literatürde yaygın kullanılan AlexNet ve VGG16 daha derin state-of-art mimarileri yanında bu çalışma için önerilen bir CNN modeli kullanılmıştır. Veri arttırımı uygulanarak gerçekleştirilen sınıflandırmalarda AlexNet ile 95.34%, VGG16 ile 98.14%, ve önerilen model ile 100% doğruluk elde edilmiştir. Elde edilen bulgular literatürdeki benzer çalışmalarda karşılaştırılmıştır. Tüm modeller ile elde edilen doğruluk oranlarının literatürdeki çalışmaların genelinden yüksek olduğu, önerilen CNN modeli ile elde edilen doğruluğun ise emsallerin tamamından yüksek olduğu sonucuna ulaşılmıştır. Ayrıca önerilen modelin daha karmaşık yapıdaki diğer modellere nazaran da yüksek başarı sergilemiş olması, daha minimal mimarilerle de yüksek başarı elde edilebileceğini ortaya koymuştur.

Keywords

• Kornea ülseri teşhisi
• evrişimsel sinir ağı
• sınıflandırma

A Deep Learning-Based Classification Study for Diagnosing Corneal Ulcers on Ocular Staining Images

Abstract

Corneal ulcer is a common disease worldwide and is one of the leading causes of corneal blindness. Diagnosis of the disease requires expertise, and the number of experienced ophthalmologists is not sufficient, especially in underdeveloped countries. For this reason, it is necessary to develop technology-based decision support systems in the diagnosis of the disease. However, the number of studies on this subject is not sufficient. In this study, CNN-based classifications were performed using corneal ulcer images obtained by an ocular staining technique, consisting of 712 samples and three classes. In addition to the AlexNet and VGG16 state-of-the-art architectures, which are widely used in the literature, a CNN model proposed for this study was used for classification. In the classifications performed by applying data augmentation, 95.34% accuracy with AlexNet, 98.14% with VGG16, and 100% accuracy with the proposed model was obtained. The findings were compared with similar studies in the literature. It was concluded that the accuracy rates obtained with all of the models used in the study were generally higher than similar studies in the literature, and the accuracy obtained with the proposed CNN model was higher than all of the peers. In addition, the success of the proposed model compared to other models with more complex structures revealed that it is not always necessary to use complex architectures for high accuracy.

Keywords

• Corneal ulcer diagnosis
• convolutional neural network
• classification

References

1. Akram, A., & Debnath, R. (2019). An Efficient Automated Corneal Ulcer Detection Method using Convolutional Neural Network. 2019 22nd International Conference on Computer and Information Technology (ICCIT), 1-6. google scholar
2. Aksoy, B. (2021). Estimation of Energy Produced in Hydroelectric Power Plant Industrial Automation Using Deep Learning and Hybrid Machine Learning Techniques. Electric Power Components and Systems, 49(3), 213-232. https://doi.org/10.1080/15325008.2021.1937401 google scholar
3. Amescua, G., Miller, D., & Alfonso, E. C. (2012). What is causing the corneal ulcer? Management strategies for unresponsive corneal ulceration. Eye, 26(2), 228-236. https://doi.org/10.1038/eye.2011.316 google scholar
4. Basak, S. K., Basak, S., Mohanta, A., & Bhowmick, A. (2005). Epidemiological and microbiological diagnosis of suppurative keratitis in Gangetic West Bengal, eastern India. Indian Journal of Ophthalmology, 53(1), 17-22. google scholar
5. Bron, A. J., Argüeso, P., Irkec, M., & Bright, F. V. (2015). Clinical staining of the ocular surface: Mechanisms and interpretations. Progress in Retinal and Eye Research, 44, 36-61. https://doi.org/10.1016/j.preteyeres.2014.10.001 google scholar
6. Chen, J., & Yuan, J. (2010). Strengthen the study of the ocular surface reconstruction. Chinese Journal of Ophthalmology, 46(1), 3-5. google scholar
7. Cohen, E. J., Laibson, P. R., Arentsen, J. J., & Clemons, C. S. (1987). Corneal ulcers associated with cosmetic extended wear soft contact lenses. Ophthalmology, 94(2), 109-114. google scholar
8. Deng, L., Lyu, J., Huang, H., Deng, Y., Yuan, J., & Tang, X. (2020). The SUSTech-SYSU dataset for automatically segmenting and classifying corneal ulcers. Scientific Data, 7(1), 23. https://doi.org/10.1038/s41597-020-0360-7 google scholar

9. Diamond, J., Leeming, J., Coombs, G., Pearman, J., Sharma, A., Illingworth, C., Crawford, G., & Easty, D. (1999). Corneal biopsy with tissue micro homogenisation for isolation of organisms in bacterial keratitis. Eye, 13(4), 545-549. google scholar
10. Garg, P., & Rao, G. N. (1999). Corneal ulcer: Diagnosis and management. Community Eye Health, 12(30), 21-23. PubMed. google scholar
11. Gross, J., Breitenbach, J., Baumgartl, H., & Buettner, R. (2021). High-performance detection of corneal ulceration using image classification with convolutional neural networks. E 54th Hawaii International Conference on System Sciences, 3416-3425. google scholar
12. Katara, R. S., Patel, N. D., & Sinha, M. (2013). A clinical microbiological study of corneal ulcer patients at western Gujarat, India. Acta Medica Iranica, 399-403. google scholar
13. Kim, J. Y., Lee, H. E., Choi, Y. H., Lee, S. J., & Jeon, J. S. (2019). CNN-based diagnosis models for canine ulcerative keratitis. Scientific Reports, 9(1), 1-7. google scholar
14. Krizhevsky, A., Sutskever, I.,& Hinton, G. E. (2017). ImageNet classification with deep convolutional neural networks. Communications of the ACM, 60(6), 84-90. google scholar
15. Li, Z., Jiang, J., Chen, K., Chen, Q., Zheng, Q., Liu, X., Weng, H., Wu, S., & Chen, W. (2021). Preventing corneal blindness caused by keratitis using artificial intelligence. Nature Communications, 12(1), 1-12. google scholar Maurice, D. M. (1957). The structure and transparency of the cornea. The Journal of Physiology, 136(2), 263. google scholar
16. Portela, H. M. B., MS Veras, R. de, Vogado, L. H. S., Leite, D., Sousa, J. A. de, Paiva, A. C. de, & Tavares, J. M. R. (2021). A Coarse to Fine Corneal Ulcer Segmentation Approach Using U-net and DexiNed in Chain. Iberoamerican Congress on Pattern Recognition, 13-23. google scholar
17. Saini, J. S., Jain, A. K., Kumar, S., Vikal, S., Pankaj, S., & Singh, S. (2003). Neural network approach to classify infective keratitis. Current Eye Research, 27(2), 111-116. google scholar
18. Sajeev, S., & Prem Senthil, M. (2021). Classifying infective keratitis using a deep learning approach. 2021 Australasian Computer Science Week Multiconference, 1-4. google scholar
19. Simonyan, K., & Zisserman, A. (2014). Very Deep ConvNets for Large-Scale Image Recognition. CoRR. google scholar
20. Song, X., Xie, L., Tan, X., Wang, Z., Yang, Y., Yuan, Y., Deng, Y., Fu, S., Xu, J., Sun, X., & others. (2014). A multi-center, cross-sectional study on the burden of infectious keratitis in China. PLoS One, 9(12), e113843. google scholar Tang, N., Liu, H., Yue, K., Li, W., & Yue, X. (2020). Automatic classification for corneal ulcer using a modified VGG network. 2020 International Conference on Artificial Intelligence and Computer Engineering (ICAICE), 120-123. google scholar
21. Teeyapan, K. (2021). Deep learning-based approach for corneal ulcer screening. The 12th International Conference on Computational Systems-Biology and Bioinformatics, 27-36. google scholar
22. Wang, T., Wang, M., Zhu, W., Wang, L., Chen, Z., Peng, Y., Shi, F., Zhou, Y., Yao, C., & Chen, X. (2021). Semi-MsST-GAN: A Semi-Supervised Segmentation Method for Corneal Ulcer Segmentation in Slit-Lamp Images. Frontiers in Neuroscience, 15. google scholar
23. Wang, T., Zhu, W., Wang, M., Chen, Z., & Chen, X. (2021). Cu-segnet: Corneal ulcer segmentation network. 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI), 1518-1521. google scholar
24. Wang, Z., Lyu, J., Luo, W., & Tang, X. (2021). Adjacent Scale Fusion and Corneal Position Embedding for Corneal Ulcer Segmentation. International Workshop on Ophthalmic Medical Image Analysis, 1-10. google scholar
25. Xu, Y., Kong, M., Xie, W., Duan, R., Fang, Z., Lin, Y., Zhu, Q., Tang, S., Wu, F., & Yao, Y.-F. (2021). Deep sequential feature learning in clinical image classification of infectious keratitis. Engineering, 7(7), 1002-1010. google scholar