Using of the machine learning methods to identify bronchopulmonary system diseases with the use of lung sounds

Abstract

This study reviews the main approaches to the analyzing of modern methods of digital processing of lung sounds. It is shown that each of the existing methods gives a definite result in solving a particular problem. However, none of the methods that were reviewed, can't be called universal and completely convenient for using in the real conditions of the hospital. Certain numerical parameters can be obtained, as a result of the work of each method. In this study it is showed that machine learning can serve as a unifying mechanism for the considered methods. A set of different parameters can be the input arguments of the classifier, which will be properly trained. As a result, the primary opinion in a convenient and accessible form can be presented to the doctor.

Artificial intelligence in the respiratory sounds analysis and computer diagnostics of bronchial asthma

“…An alternative approach could be the following: to develop diagnostic methods based on machine learning and deep machine learning techniques. Recently, the studies of the possibility of applying such methods in various fields of medicine have been actively carried out (see [12,13,[17][18][19][20][21] and references therein). At present, there is a large number of machine learning methods that allows creating decision-making algorithms suitable for a given situation and able to solve various tasks, including medical ones (e.g., classification of various types of cancer [18], analysis of the effectiveness of dialysis [18], diagnosis of pulmonary diseases [12,13,20,36]).…”

“…Recently, the studies of the possibility of applying such methods in various fields of medicine have been actively carried out (see [12,13,[17][18][19][20][21] and references therein). At present, there is a large number of machine learning methods that allows creating decision-making algorithms suitable for a given situation and able to solve various tasks, including medical ones (e.g., classification of various types of cancer [18], analysis of the effectiveness of dialysis [18], diagnosis of pulmonary diseases [12,13,20,36]). Machine learning methods are proposed to be widely used, in particular, in pulmonology: for sorting patients with chronic obstructive pulmonary disease [21], for automatic classification of respiratory sound spectra in patients with pulmonary diseases [22], for diagnosis of bronchial asthma and chronic obstructive pulmonary disease [23] and ranking of clinical assessments in children with asthma [24].…”

“…Computer-aided methods for diagnosing pulmonary diseases are devoid of subjectivity and allow analyzing changes in respiratory sounds, which cannot be detected by the human ear. Changes in the characteristics of the airways caused by various diseases lead to appearance of additional pathological noises, the analysis of which is used in development of computer-aided methods for the diagnostics of pulmonary diseases [6,7,12]. Respiratory sounds, accompanying broncho-obstructive syndrome can be characterized by the appearance / amplification of a set of periodic waves with a fundamental frequency in the range from 100 Hz to 2500 Hz.…”

“…), communication facilities, and software makes it is possible to develop new methods of computer-aided diagnostics based on the analysis of patients' breathing sounds. These methods include machine learning methods and deep machine learning methods [12,13]. However, the development of such diagnostic methods requires the analysis of a large database with the maximum possible number of patients and volunteers.…”

Association Between The Respiratory Organ’s Morbidity Of Adult Population And Territorial Aridity

“…From the point of view of preventive and clinical medicine the assessment of the impact of climate on the incidence of respiratory diseases is very relevant, but to date the problem has not been fully resolved. The analysis of respiratory diseases and the contribution of individual factors to their formation is successfully implemented using information technologies (Schweitzer et al, 2018), such as machine learning technology (Santamouris et al, 2020), evolutionary algorithms method (Chekhovych et al, 2018), neural network method (Dwivedi, 2018), digital platform technologies and web portals (Fernández et al, 2022), telemedicine (Mosnaim et al, 2021).…”