Treinamento de Classificadores de Aprendizagem de Máquina para o Diagnóstico do Glaucoma com o Uso de Dados de Perimetria Automatizada Padrão (SAP)

• Bruno F. Hashimoto USP
• Leonardo S. Shigueoka UNICAMP
• Vital P. Costa UNICAMP
• Edson S. Gomi USP

DOI: https://doi.org/10.5753/sbcas.2021.16117

Abstract

Glaucoma is an optical neuropathy that causes damage to the optic nerve and consequent loss of the visual field. This research project investigates the performance of Machine Learning Classifiers in the diagnosis of glaucoma using data from the SAP exam (Standard Automated Perimetry). Due to the dataset with real patient data being small, a synthetic dataset was generated to increase the amount of data available for the training of the classifiers. The classifiers Random Forest, Gradient Boosting and DNN (Deep Neural Network) were tested. The results obtained show that these classifiers present acceptable performance for the diagnosis of glaucoma, all classifiers presented AUC above 83,9% and accuracy above 72,0%.

References

Asaoka, R., Murata, H., Iwase, A., and Araie, M. (2016). Detecting preperimetric glaucoma with standard automated perimetry using a deep learning classifier. Ophthalmology, 123(9):1974 – 1980.

Azuara-Blanco, A., Costa, V., and Wilson, R. (2001). Handbook of Glaucoma. Taylor & Francis.

Breiman, L. (2001). Random forests. Machine learning, 45(1):5–32.

Chollet, F. (2017). Deep Learning with Python. Manning Publications Co., USA, 1st edition.

Friedman, J. H. (2001). Greedy function approximation: a gradient boosting machine. The Annals of Statistics, 29(5):1189–1232.

Hinton, G. E. and Salakhutdinov, R. R. (2006). Reducing the dimensionality of data with neural networks. science, 313(5786):504–507.

Liu, J. H. K., Zhang, X., Kripke, D. F., and Weinreb, R. N. (2003). Twenty-four-Hour Intraocular Pressure Pattern Associated with Early Glaucomatous Changes. Investigative Ophthalmology and Visual Science, 44(4):1586–1590.

Mandrekar, J. N. (2010). Receiver operating characteristic curve in diagnostic test assessment. Journal of Thoracic Oncology, 5(9):1315–1316.

Shigueoka, L. S., Vasconcellos, J. P. C. d., Schimiti, R. B., Reis, A. S. C., Oliveira, G. O. d., Gomi, E. S., Vianna, J. A. R., Lisboa, R. D. d. R., Medeiros, F. A., and Costa, V. P. (2018). Automated algorithms combining structure and function outperform general ophthalmologists in diagnosing glaucoma. PloS one, 13(12):e0207784.

Shorten, C. and Khoshgoftaar, T. M. (2019). A survey on image data augmentation for deep learning. Journal of Big Data, 6(1):1–48.

Vincent, P., Larochelle, H., Lajoie, I., Bengio, Y., Manzagol, P.-A., and Bottou, L. (2010). Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion. Journal of machine learning research, 11(12):3371– 3408.