Application of Neural Networks for Differential Diagnosis of Pulmonary Pathologies Based on X-Ray Images

This work is devoted to studying the possibility of creating intelligent automated systems for differential diagnosis of pulmonary diseases based on the identification of pathological structures on X-ray images of the thoracic cavity organs (TCO) using neural network technologies. A brief analysis of modern diagnostic techniques is presented; a description of the proposed algorithm for determining the type of lung tissue pathologies used in the visual analysis of X-ray images and based on the identification of the main radiological syndromes, as well as on the evaluation of the quantitative characteristics of differential X-ray diagnostics is given. By the example of classification of radiographs of healthy and tuberculosis patients, the effectiveness of the use of neural network technologies in the computer diagnosis of lung diseases is demonstrated. The studies have been carried out using a publicly available database of X-ray images of thoracic cavity organs containing 3500 images of healthy people and 3500 images of sick people.