UDC 621.397
MODIFICATION OF THE ALGORITHM FOR CLASSIFYING NOISES ON IMAGES USING A BACK PROPAGATION NEURAL NETWORK
V. T. Dmitriev, Dr. Sc. (Tech.), associate professor, department of radio control and communication, Head of the Department, RSREU, Ryazan, Russia;
orcid.org/0000-0001-5521-6886, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
E. S. Cherentaeva, post-graduate student, RSREU, Ryazan, Russia;
orcid.org/0000-0002-5580-3033, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
M. A. Cherentaev, post-graduate student, RSREU, Ryazan, Russia;
orcid.org/0000-0002-0282-7250, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
A modification of image noise classification algorithm using back propagation neural network has been developed. The aim of the work is software implementation of noise classification neural network with the modification of preliminary processing algorithm with an increase in the number of neural network layers, as well as testing the resulting classifier on a set of real satellite images. An algorithm for preliminary data processing for training the neural network has been developed. The probabilities of correctly determining the type of noise have been obtained and compared with the values for other known classification methods. The authors have shown that the modified neural network provides better classification for multiplicative and Gaussian noise compared to known algorithms.
Key words: : : image noise classification, noise models, image denoising, image filtering, statistical moment, multilayer neural network, back propagation method, gradient descent