This article is cited in 10 papers

IMAGE PROCESSING, PATTERN RECOGNITION

Neural network application for semantic segmentation of fundus

R. A. Paringer^ab, A. V. Mukhin^a, N. Yu. Ilyasova^ab, N. S. Demin<sup>ab

a</sup> Samara National Research University
^b Image Processing Systems Institute of the RAS - Branch of the FSRC "Crystallography and Photonics" RAS, Samara, Russia, Samara

Abstract: Advances in the neural networks have brought revolution in many areas, especially those related to image processing and analysis. The most complex is a task of analyzing biomedical data due to a limited number of samples, imbalanced classes, and low-quality labelling. In this paper, we look into the possibility of using neural networks when solving a task of semantic segmentation of fundus. The applicability of the neural networks is evaluated through a comparison of image segmentation results with those obtained using textural features. The neural networks are found to be more accurate than the textural features both in terms of precision ($\sim25\%$) and recall ($\sim50\%$). Neural networks can be applied in biomedical image segmentation in combination with data balancing algorithms and data augmentation techniques.

Keywords: convolution, neural network, convolutional network, segmentation, fundus

Received: 09.07.2021
Accepted: 25.11.2021

DOI: 10.18287/2412-6179-CO-1010