Separation of cardiorespiratory sounds using time-frequency masking and sparsity

Abstract

Listening to cardiac and respiratory sounds called as auscultation is a non-invasive medical procedure, which provides useful information about the behavior of the heart and the lung. Cardiac and respiratory sounds interfere with each other as well as with other sounds like snore, speech or traffic noise, which compromises the effectiveness of auscultation. This paper addresses the problem of auscultation in complex auditory environments, inspired by the coincidence detection model which suggests sound localization via estimating interaural level difference and interaural time difference. The proposed method, exploits the sparsity of cardiac and respiratory sounds and makes use of a degenerate unmixing estimation technique (DUET), which uses only two observations to recover an arbitrary number of sources, which suits well in scenarios where the number of sources can vary. The DUET approach uses timefrequency analysis to produce a two dimensional histogram of attenuation-delay estimates, where peaks in the histogram indicate the sources in a mixture. A mask is computed using attenuation-delay mixing parameters to recover the original sources. It is shown that excellent time-frequency masks exist for cardiac and respiratory sounds. The performance of the proposed method is demonstrated through a series of experiments using real data, exhibiting superior source recovery than previous techniques.

Monaural cardiopulmonary sound separation via complex-valued deep autoencoder and cyclostationarity

“…The adaptive linear enhancement (ALE) was used for cardiopulmonary sound separation in [16], where an introduced delay caused decorrelation between the noise components of the input signal and its delayed filtered version. Based on based on the assumption that cardiac and lung sound signals are sparse, the idea of time-frequency masking was used to address this issue [17]. Nonnegative matrix factorization (NMF) was first introduced in [18] to separate cardiac and lung sounds.…”

Source separation for single channel thoracic cardio-respiratory sounds applying Non-negative Matrix Factorization (NMF) using a focused strategy on heart sound positions

“…1 Physicians use auscultation as a simple, non-invasive and low-cost method to obtain relevant information about the state of internal organs. 2,3 Lung sounds are produced by a turbulent flow o f a ir w ithin t he r espiratory t ract d uring i nhalation and exhalation processes, [4][5][6] mainly in the bronchi and trachea. 1 This flow p ropagates i n f orm o f s ound through the lung tissues which can be heard on the chest wall.…”