Using a Neural Network to Separate Lungs in X-Ray Images

Abstract

An auxiliary neural network is developed to solve the problem of preliminarily processing X-ray images to distinguish areas with lung and spine, and to mask unnecessary areas. Images processed with the neural network are used to study the main neural network for detecting pathology in lungs. A comparative analysis is performed of the quality of studying the main neural network on the basis of preliminarily processed images and initial ones.