Classificação de Estágios de Glaucoma Utilizando Volumes OCT e CNN3D

• Matheus Levy UFMA
• Marcos Melo Ferreira UFMA
• Geraldo Braz Júnior UFMA
• Marcus Vinicius Silva Lima de Oliveira UFMA
• Gabriel Costa UFMA
• Mario Freitas UFMA
• Luis Rivero UFMA

DOI: https://doi.org/10.5753/sbcas.2022.222659

Abstract

Glaucoma is one of the diseases responsible for blindness. The detection of this disease at an early stage can help patients have treatment to improve their quality of life. This study outlined the use of a method based on Deep Learning to make the classification of glaucoma stages in early, mid/advanced and without glaucoma. For this, we will use GAMMA database, composed of 100 pairs of images, color fundus photographys and three-dimensionals Optical Coherence Tomography (OCT). To classify this type of three-dimensional data, we propose the use of a 3D CNN based on the principles of DenseNet and used Transfer Learning. The best result was 84% of ACC with 2 Classes and 67% with 3 Classes. Even with limited results, the technique proves to be a good basis for future improvement work.

References

Albawi, S., Mohammed, T. A., and Al-Zawi, S. (2017). Understanding of a convolutional neural network. pages 1-6. Ieee.

An, G., Omodaka, K., Hashimoto, K., Tsuda, S., Shiga, Y., Takada, N., Kikawa, T., Yokota, H., Akiba, M., and Nakazawa, T. (2019). Glaucoma diagnosis with machine learning based on optical coherence tomography and color fundus images. Journal of healthcare engineering, 2019.

Araújo, F., Carneiro, A., Silva, R., Medeiros, F., and Ushizima, D. (2017). Redes neurais convolucionais com tensorflow:teoria e prática. 1:382-406.

Bourne, R. R. et al. (2021). Causes of blindness and vision impairment in 2020 and trends over 30 years, and prevalence of avoidable blindness in relation to vision 2020: the right to sight: an analysis for the global burden of disease study. The Lancet Global Health, 9(2):e144-e160.

Fernandes, A. G., Martins, C., Lima, A. C., Junior, G., de Almeida, J. D., and de Paiva, A. (2019). Meta aprendizagem de extração de características aplicada ao diagnóstico de glaucoma. In Anais do XIX Simpósio Brasileiro de Computação Aplicada à Saúde, pages 342-347, Porto Alegre, RS, Brasil. SBC.

Fricke, T. R., Tahhan, N., Resnikoff, S., Papas, E., Burnett, A., Ho, S. M., Naduvilath, T., and Naidoo, K. S. (2018). Global prevalence of presbyopia and vision impairment from uncorrected presbyopia: systematic review, meta-analysis, and modelling. Ophthalmology, 125(10):1492-1499.

Fu, H., Baskaran, M., Xu, Y., Lin, S., Wong, D. W. K., Liu, J., Tun, T. A., Mahesh, M., Perera, S. A., and Aung, T. (2019a). A deep learning system for automated angleclosure detection in anterior segment optical coherence tomography images. American journal of ophthalmology, 203:37-45.

Goodfellow, I., Bengio, Y., and Courville, A. (2016). Deep Learning. MIT Press, Cambridge. http://www.deeplearningbook.org.

Li, J., Chen, Y., Cai, L., Davidson, I., and Ji, S. (2017). Dense transformer networks. arXiv preprint arXiv:1705.08881.

Ma, J., Lv, B., Li, Y., Fan, P., Zhao, X., Yuan, H., and Zhang, Y. (2021). Multimodal primary open angle glaucoma early diagnosing program based on clinical process.

Maetschke, S., Antony, B., Ishikawa, H., Wollstein, G., Schuman, J., and Garnavi, R. (2019). A feature agnostic approach for glaucoma detection in oct volumes. PloS one, 14(7):e0219126.

Martins, C., Santos, R. D., Almeida, J. D., Júnior, G. B., and Teixeira, J. A. (2021). Detecção de patologias oculares em imagens de reflexo vermelho utilizando descritores de cor. In Anais do XXI Simpósio Brasileiro de Computação Aplicada à Saúde, pages 346-357, Porto Alegre, RS, Brasil. SBC.

Mehta, P., Petersen, C. A., Wen, J. C., Banitt, M. R., Chen, P. P., Bojikian, K. D., Egan, C., Lee, S.-I., Balazinska, M., Lee, A. Y., et al. (2021). Automated detection of glaucoma with interpretable machine learning using clinical data and multimodal retinal images. American Journal of Ophthalmology, 231:154-169.

Moreira, J. M., Almeida, J. D., Júnior, G. B., and Paiva, A. (2021). Detecção de glaucoma usando redes em cápsula. In Anais do XXI Simpósio Brasileiro de Computação Aplicada à Saúde, pages 188-199, Porto Alegre, RS, Brasil. SBC.

Muhammad, H., Fuchs, T. J., De Cuir, N., De Moraes, C. G., Blumberg, D. M., Liebmann, J. M., Ritch, R., and Hood, D. C. (2017). Hybrid deep learning on single wide-field optical coherence tomography scans accurately classifies glaucoma suspects. Journal of glaucoma, 26(12):1086.

Quigley, H. A. and Broman, A. T. (2006). The number of people with glaucoma worldwide in 2010 and 2020. British journal of ophthalmology, 90(3):262-267.

Sarhan, A., Rokne, J., and Alhajj, R. (2019). Glaucoma detection using image processing techniques: A literature review. Computerized Medical Imaging and Graphics, page 101657.

SBG (2009). 3º consenso de glaucoma primário de ângulo aberto. https://www.sbglaucoma.org.br/wp-content/uploads/2020/06/consenso03-v2.pdf. Acessado em: 12-Janeiro-2022.

Solovyev, R., Kalinin, A. A., and Gabruseva, T. (2022). 3d convolutional neural networks for stalled brain capillary detection. Computers in Biology and Medicine, 141:105089.

WHO (2019). World report on vision. https://www.who.int/publications/i/item/9789241516570. Acessado em: 12-Janeiro-2022.

Wu, J., Fang, H., Li, F., Fu, H., Lin, F., Li, J., Huang, L., Yu, Q., Song, S., Xu, X., et al. (2022). Gamma challenge: Glaucoma grading from multi-modality images. arXiv preprint arXiv:2202.06511.

Xiong, J., Li, F., Song, D., Tang, G., He, J., Gao, K., Zhang, H., Cheng, W., Song, Y., Lin, F., et al. (2021). Multimodal machine learning using visual fields and peripapillary circular oct scans in detection of glaucomatous optic neuropathy. Ophthalmology.

Xu, B. Y., Chiang, M., Chaudhary, S., Kulkarni, S., Pardeshi, A. A., and Varma, R. (2019). Deep learning classifiers for automated detection of gonioscopic angle closure based on anterior segment oct images. American journal of ophthalmology, 208:273-280.