Skip to Main Content
This paper addresses the problem of heart sounds (HS) localization from single channel respiratory sounds (RS) recordings by applying wavelet-based localization scheme. After a wavelet-based multiscale decomposition of the noisy signal, HS contaminated segments are localized in the noisy RS signal based on the cumulative sums of likelihood ratios capturing the dynamic behaviour of the signal. Quantitative evaluation of the localized HS segments for various types of simulated data has been performed. The comparisons between the estimated boundaries of the localized HS segments and the actual segment boundaries of the synchronized pure HS signals show the proposed method is able to localize the HS segments accurately in an automatic way. Also, the test results on real RS recordings in terms of the detection accuracy show the promising performance by the proposed method.