Capabilities of computer analysis of breath sounds in patients with COVID-19

Abstract

Objective. To develop methods for a rapid distance computer diagnosis of COVID-19 based on the analysis of breath sounds. It is known that changes in breath sounds can be the indicators of respiratory organs diseases. Computer analysis of these sounds can indicate their typical changes caused by COVID-19, and can be used for a rapid preliminary diagnosis of this disease.
 Materials and methods. The method of fast Fourier transform (FFT) was used for computer analysis of breath sounds, recorded near the mouth of 14 COVID-19 patients (aged 1880 years) and 17 healthy volunteers (aged 548 years). The frequency of breath sound records ranged from 44 to 96 kHz. Unlike the conventional methods of computer analysis for diagnosis of diseases based on respiratory sound studying, we offer to test a high-frequency part of FFT (20006000 kHz).
 Results. While comparing the breath sound FFT in patients and healthy volunteers, we developed the methods for COVID-19 computer diagnosis and determined the numerical criteria in patients and healthy persons. These criteria do not depend on sex and age of the examined persons.
 Conclusions. The offered computer methods based on the analysis of breath sound FFT in patients and volunteers permit to diagnose COVID -19 with relatively high diagnostic parameters. These methods can be used in development of noninvasive means for preliminary self-express diagnosis of COVID-19.

Application of deep learning for bronchial asthma diagnostics using respiratory sound recordings

“…In pulmonary diseases, pathological processes and the physiological changes provoked by them are reflected in the character of respiratory sounds. The computer-assisted diagnostics methods developed previously that utilize the Fourier spectrum analysis or other sound wave metrics attempt to link disease diagnostics with the analysis of biological, physiological, and clinical processes in the patient’s body ( Furman et al, 2014 ; Furman et al, 2015 ; Furman et al, 2021 ; Furman et al, 2020 ; Gelman et al, 2021 ). Machine learning and deep learning methods are designed to divide data into certain categories according to identified common attributes that cannot always be clearly matched to body processes, and to statistically analyze the signals.…”

“…The files containing respiratory sound recordings created with phones with both internal and external microphones were transferred to a cloud storage to create an anonymous database. Computer recordings were done using developed computer systems for respiratory sound recording ( Furman et al, 2014 ; Furman et al, 2015 ; Furman et al, 2021 ; Furman et al, 2020 ). These computer systems contain external microphones, electronic phonendoscopes, and computer sound cards, as well as Adobe Audition software.…”

“…However, multiple different lung diseases can be accompanied by the same pathological respiratory sounds ( GINA, 2022 ; Furman et al, 2020 ; Sovijarvi et al, 2000 ). Computer-assisted analysis of breath sounds can become an additional method of preclinical diagnostics of lung diseases, including BA in children; it can also be an instrument for objective express diagnostics of the disease and remote monitoring of the state of patient’s health, becoming a component of telemedicine ( Furman et al, 2014 ; Furman et al, 2015 ; Furman et al, 2021 ; Furman et al, 2020 ; Chekhovych et al, 2018 and references therein). Computer-assisted methods of lung disease diagnostics are free of subjectivity and enable the ability to analyze changes in respiratory sounds, which cannot be detected by the human ear.…”

Computer-Aided Detection of Respiratory Sounds in Bronchial Asthma Patients Based on Machine Learning Method

“…Files recorded on phones using the built-in and/or external microphones were sent to a “cloud” to create an anonymous database. Computer-aided records were made using systems for recording respiratory sounds [ 4 , 5 , 7 , 11 ] with external microphones, electronic phonendoscopes, and computer sound cards; with the Adobe Audition audio editor. All systems exhibited high amplitude-frequency linearity in the 100 to 3000 Hz frequency range.…”

“…Computer-aided analysis of respiratory sounds may complement the screening diagnosis of pulmonary diseases including BA [4][5][6][7][8]. Computer-aided detection methods are able to analyze changes of respiratory sounds that cannot be detected by a human ear.…”