Subject independent identification of breath sounds components using multiple classifiers

Abstract

Breath sounds have been shown very valuable for diagnosis of obstructive sleep apnea. In this study, we present a subject independent method for automatic classification of breath and related sounds during sleep. An experienced operator manually labelled segments of breath sounds from 11 sleeping subjects as: inspiration, expiration, inspiratory snoring, expiratory snoring, wheezing, other noise, and non-audible. Ten features were extracted and fed into 3 different classifiers: naïve Bayes, Support Vector Machine, and Random Forest. Leave-one-out method was used in which data from each subject, in turn, is evaluated using models trained with all other subject. Mean accuracy for concurrent classification of all 7 classes reached 85.4%. Mean accuracy for separating data into 2 classes, snoring and non-snoring, reached 97.8%. To our knowledge, these are the highest accuracies achieved in automatic classification of all breath sounds components concurrently and for snoring, in a subject independent model.

Objective Relationship Between Sleep Apnea and Frequency of Snoring Assessed by Machine Learning

“…Snores were automatically detected using an algorithm for breath sounds classification that has been previously described and validated against the human ear. 17 Briefly, in that study, a human operator, who was blinded to the machine score, manually annotated audio of breath sounds based on this definition: snoring is a harsh, low-pitched sound produced by tissue vibration during passage of air through a partially collapsed upper airway during sleep. 23,24 Separately, audio files were segmented into 64 ms, from each 10 acoustic variables (features) that were calculated to be used by the machine learning algorithm.…”

“…To overcome these limitations, we have developed a system to detect individual snores based on the acoustic nature of the breath sounds regardless of its amplitude solely. 17 This system is capable of accurately identifying individual snores, which are then quantified in order to evaluate their relationship to sleep apnea (both OSA and CSA). Machine learning algorithms have been deployed by other researchers for identifying snoring and examining snoring acoustic features in snorers with and without OSA, in a small set of participants.…”

Destek Vektör Makineleri Kullanarak Uyku Seslerinin Çoklu Sınıflandırılması

“…Bu senkronizasyon işleminin doğru şekilde yapılamaması, uzman doktorun teşhis koyma sürecinde hata yapmasına sebebiyet verebilir. İkinci problem ise, ortalama uyku süresinin 6-8 saat olmasından dolayı, uzman doktorun ses kaydının tamamını dinleyerek teşhis koyabilmesinin pratikte zor olmasıdır (Alshaer et al, 2014).…”

“…Nefes alma, nefes verme, nefes alırken horlama, nefes verirken horlama, hırıltı, diğer sesler ve duyulamayan seslerden oluşan yedi farklı ses tipinin incelendiği diğer bir çalışmada elde edilen öznitelikler kullanılarak üç farklı sınıflandırıcı kullanılmıştır. Sonuçlara göre; uyku sesleri konusunda ikiden fazla parametrenin sınıflandırılması durumunda, en yüksek doğruluk oranının % 85.4 olarak elde edildiği bildirilmiştir (Alshaer et al, 2014).…”

Real-time Identification of Respiratory Movements through a Microphone

“…Methods belonging to Artificial Intelligence disciplines are used to identify respiratory events. K-Nearest Neighbour [OUD, M. & DOOIJES, E.H., 1996], Naive Bayes [ALSHAER, H. et al, 2014], Support Vector Machines [ALSHAER, H. et al, 2014], Random Forest [ALSHAER, H. et al, 2014], Neural Networks [YUAN, K. et al, 2011;PATEL, U., 2011;MASON, L., 2002] and Hidden Markov Models [SNIDER, B.R. & KAIN, A., 2013] are common techniques that have been proven to achieve good results.…”