A Simple Gaussian Filter can Protect Images of Noise Injection in a Classification Task with Deep Neural Networks

Authors

  • Christian Ruiz Hernández Computer Science Department at the Center of Research and Advanced Studies (CINVESTAV)
  • Luis Gerardo De la Fraga Computer Science Department at the Center of Research and Advanced Studies (CINVESTAV) https://orcid.org/0000-0002-9373-9837

DOI:

https://doi.org/10.66131/JDSC13202611-22

Keywords:

Convolutional Neural Networks, Deep Neural Networks, Noise Attack, Denoising

Abstract

We study the classification of a set of images under additive Gaussian noise. We observe that the very simple denoising Gaussian
filter obtain the best results for a classification problem using a Deep Convolutional Network. We were able to generate an attack
of noise injection to change the class of two images, and the correct class can be obtained applying also the very simple Gaussian
filter.

References

B. García-Morgado. “Análisis comparativo de algoritmos de clasificación de imágenes basados en aprendizaje automático tradicional y aprendizaje profundo”. MA thesis. Cinvestav, 2024.

J. Su, D.V. Vargas, and K. Sakurai. “One pixel attack for fooling deep neural networks”. In: IEEE Transactions on Evolutionary Computation 23.5 (2019), pp. 828–841.

Luana Clare, Alexandra Marques, and João Correia. “A Comparative Analysis of Evolutionary Adversarial One-Pixel Attacks”. In: Applications of Evolutionary Computation. Ed. by Stephen Smith, João Correia, and Christian Cintrano. Cham: Springer Nature Switzerland, 2024, pp. 147–162. isbn: 978-3-031-56855-8.

M.W. Kusk and S. Lysdahlgaard. “The effect of Gaussian noise on pneumonia detection on chest radiographs, using convolutional neural networks”. In: Radiography 29.1 (2023), pp. 38–43. doi: 10.1016/j.radi.2022.09.011. url: https://doi.org/10.1016/j.radi.2022.09.011.

21

Teresa Kwamboka Abuya, Richard Maina Rimiru, and George Onyango Okeyo. “An Image Denoising Technique Using Wavelet-Anisotropic Gaussian Filter-Based Denoising Convolutional Neural Network for CT Images”. In: Applied Sciences 13.21 (2023). issn: 2076-3417. doi: 10.3390/app132112069. url: https://www.mdpi.com/2076-3417/13/21/12069.

Y. Bi, B. Xue, and M. Zhang. “Genetic Programming-Based Evolutionary Deep Learning for Data-Efficient Image Classification”. In: IEEE Transactions on Evolutionary Computation 28.2 (2024), pp. 307–322. doi: 10.1109/TEVC.2022.3214503.

Y. Bi, M. Zhang, and B. Xue. “Genetic Programming for Automatic Global and Local Feature Extraction to Image Classification”. In: 2018 IEEE Congress on Evolutionary Computation (CEC). 2018, pp. 1–8. doi: 10.1109/CEC.2018.8477911.

G. Olague. Evolutionary Computer Vision. Springer Science+Business Media, Jan. 2016. doi: 10.1007/978-3-662-43693-6. (Visited on 11/12/2024).

S.M. Moosavi-Dezfooli, A. Fawzi, and P. Frossard. “DeepFool: A Simple and Accurate Method to Fool Deep Neural Networks”. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). June 2016, pp. 2574–2582.

Additional Files

Published

2026-10-30

How to Cite

Ruiz Hernández, C., & De la Fraga, L. G. (2026). A Simple Gaussian Filter can Protect Images of Noise Injection in a Classification Task with Deep Neural Networks. Journal of Dynamical Systems and Complexity , 1(3), 11–22. https://doi.org/10.66131/JDSC13202611-22

Issue

Vol. 1 No. 3 (2026): Selected papers 2026

Section

Artificial Intelligence and Cybersecurity

License

Copyright (c) 2026 Christian Ruiz Hernández, Luis Gerardo De la Fraga

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.