Detecção automática de doenças da visão em imagens de reflexo vermelho utilizando Deep Features e Ensemble

• Matheus Henrique A. Nunes UFMA
• João Dallyson S. Almeida UFMA
• Italo Francyles S. da Silva UFMA
• Geraldo Braz Júnior UFMA

DOI: https://doi.org/10.5753/sbcas.2023.229638

Abstract

Brückner’s Test, popularly known as red reflex exam, is a simple and painless diagnosis method aimed at early detection of vision related diseases. Observing retinal red reflex with an direct ophthalmoscope, internal structural features are identified and can alert to a compromised eye health. The methodology combines characteristic descriptors based on deep learning and classifiers to identify the presence of pathologies in red reflex images, and studies highlighted the use of the convolucional neural network DeepLoc and an ensemble combining Linear Regression, Random Forest and SVM classifiers, achieving 93.20% of accuracy, sensibility of 84.50% and specificity of 93.20%.

References

Aguiar, A. S. C. d. (2010). Validação de tecnologia para avaliação do teste do reflexo vermelho. Dissertação (Mestrado em Enfermagem) Faculdade de Farmácia, Odontologia e Enfermagem. Universidade Federal do Ceará, Fortaleza.

Almagro Armenteros, J. J., Sønderby, C. K., Sønderby, S. K., Nielsen, H., and Winther, O. (2017). Deeploc: prediction of protein subcellular localization using deep learning. Bioinformatics, 33(21):3387–3395.

CBO (2019). Atenção primária à saúde no brasil. [link].

Cruz, A. (2008). Atenção primária à saúde no brasil. [link].

Demšar, J., Curk, T., Erjavec, A., Gorup, Č., Hočevar, T., Milutinovič, M., Možina, M., Polajnar, M., Toplak, M., Starič, A., et al. (2013). Orange: data mining toolbox in python. the Journal of machine Learning research, 14(1):2349–2353.

Godec, P., Pančur, M., Ilenič, N., Čopar, A., Stražar, M., Erjavec, A., Pretnar, A., Demšar, J., Starič, A., Toplak, M., et al. (2019). Democratized image analytics by visual programming through integration of deep models and small-scale machine learning. Nature communications, 10(1):4551.

Hastie, T., Tibshirani, R., and Friedman, J. (2009). The elements of statistical learning: data mining, inference, and prediction. Springer.

Iandola, F. N., Han, S., Moskewicz, M. W., Ashraf, K., Dally, W. J., and Keutzer, K. (2016). Squeezenet: Alexnet-level accuracy with 50x fewer parameters and¡ 0.5 mb model size. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 4856–4864.

Khedekar, A., Devarajan, B., Ramasamy, K., Muthukkaruppan, V., and Kim, U. (2019). Smartphone-based application improves the detection of retinoblastoma. Eye, 33:1.

Kriangsakdachai, S., Ayudhya, S. P. N., Kusakunniran, W., Ayudhya, W. D. N., Chantrasagul, C., Manasboonpermpool, R., Sathianvichitr, K., Sangsre, P., and Surachatkumtonekul, T. (2022). Anomaly detection in red reflex images using deep learning approaches. In TENCON 2022-2022 IEEE Region 10 Conference (TENCON), pages 1–6. IEEE.

Ma, J., Jiang, X., Fan, A., Jiang, J., and Yan, J. (2021). Image matching from handcrafted to deep features: A survey. International Journal of Computer Vision, 129(1):23–79.

Martins, C. M. d. S., de Santos, R. D. A., Almeida, J. D. S. d., Junior, G. B., and Teixeira, J. A. M. (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. SBC.

Pinto, A. M., Almeida, J. D. S., Braz Júnior, G., and Silva, I. F. S. (2020). Detecção de patologias da visão em imagens de reflexo vermelho utilizando deep learning. In Anais da VIII Jornada de Informática do Maranhão (eJIM).

Quinlan, J. R. (1986). Induction of decision trees. In Machine learning, volume 1, pages 81–106. Springer.

Rivas-Perea, P., Baker, E., Hamerly, G., and Shaw, B. F. (2014). Detection of leukocoria using a soft fusion of expert classifiers under non-clinical settings. BMC ophthalmology, 14:1–15.

Silva, I. F., Almeida, J. D., Teixeira, J. A., Junior, G. B., and de Paiva, A. C. (2018). Teste automático de brückner basedo em imagens. In Anais do XVIII Simpósio Brasileiro de Computação Aplicada à Saúde. SBC.

Silva, I. F. S. d. (2019). Detecção automática da presença de patologia na visão baseada em imagens do teste de brückner. Master’s thesis, Universidade Federal do Maranhão.

Simonyan, K. and Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.

Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., and Wojna, Z. (2016). Rethinking the inception architecture for computer vision. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 2818–2826.

Tamura, M. and Teixeira, L. F. (2009). Leukocoria and the red reflex test. Einstein (São Paulo), 7.

Tan, M. and Le, Q. V. (2019). Efficientnet: Rethinking model scaling for convolutional neural networks. pages 6105–6114. PMLR.

Zebari, R., Abdulazeez, A., Zeebaree, D., Zebari, D., and Saeed, J. (2020). A comprehensive review of dimensionality reduction techniques for feature selection and feature extraction. Journal of Applied Science and Technology Trends, 1(2):56–70.

Zoph, B., Vasudevan, V., Shlens, J., and Le, Q. V. (2018). Learning transferable architectures for scalable image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 8697–8710.