Применение полиспектрального анализа для определения диагностических признаков в звуках дыхания больных ХОБЛ

Abstract

In the paper a method for analyzing breath sounds of healthy people and patients with chronic obstructive lung disease on the basis of higher-order statistics was proposed. Widely used in medicine and is an electronic auscultation, which allows you to identify and objectify the typical diagnostic signs lung disease. The complex nature of the breath sounds is causing the application to the analysis of their methods of higher-order statistics. This is of interest not only to the spectral components of the respiratory sounds, as well as the phase components. Iterative method of sound processing was developed and it’s result is classification of the studied sounds by category "healthy" and "COLD patients» with a high probability. The proposed methodology is based on the calculations of bicoherence functions and skewness coefficients. It is shown that this method is an informative, highly accurate and can be a useful tool in the diagnosis of bronchopulmonary diseases.