Special Issue on Acoustic Sensors and Their Applications (Vol. 1).

Acoustic sensors have been in commercial use for more than 60 years [...].

A New Acoustical Autonomous Method for Identifying Endangered Whale Calls: A Case Study of Blue Whale and Fin Whale.

In this paper, we study to improve acoustical methods to identify endangered whale calls with emphasis on the blue whale (Balaenoptera musculus) and fin whale (Balaenoptera physalus). A promising method using wavelet scattering transform and deep learning is proposed here to detect/classify the whale calls quite precisely in the increasingly noisy ocean with a small dataset. The performances shown in terms of classification accuracy (>97%) demonstrate the efficiency of the proposed method which outperforms the relevant state-of-the-art methods. In this way, passive acoustic technology can be enhanced to monitor endangered whale calls. Efficient tracking of their numbers, migration paths and habitat become vital to whale conservation by lowering the number of preventable injuries and deaths while making progress in their recovery.

A Fully-Automated Method to Evaluate Coronavirus Disease Progression with COVID-19 Cough Sounds Using Minimal Phase Information.

This paper focuses on an important issue of disease progression of COVID-19 (coronavirus disease 2019) through processing COVID-19 cough sounds by proposing a fully-automated method. The new method is based on time-domain exploiting only phase 1 data which is always available for any cough events. The proposed approach generates plausible click sequences consist of clicks for various cough samples from covid-19 patients. The click sequence, which is extracted from the phase slope function of an input signal, is used to calculate inter-click intervals (ICIs), and thereby a scoring index (SI) is derived based on coefficient of variation(CV) of the extracted ICIs. Moreover, probability density function (pdf) of the output click sequence is obtained. The method does not need to adjust any parameters. The experimental results achieved from real-recorded COVID-19 cough data using the medically annotated Novel Coronavirus Cough Database (NoCoCoDa) reveal that the proposed time-domain method can be a very useful tool for automatic cough sound processing to determine the disease progression of coronavirus patients.

Signal feature extraction by multi-scale PCA and its application to respiratory sound classification.

Respiratory sound (RS) signals carry significant information about the underlying functioning of the pulmonary system by the presence of adventitious sounds. Although many studies have addressed the problem of pathological RS classification, only a limited number of scientific works have focused in multi-scale analysis. This paper proposes a new signal classification scheme for various types of RS based on multi-scale principal component analysis as a signal enhancement and feature extraction method to capture major variability of Fourier power spectra of the signal. Since we classify RS signals in a high dimensional feature subspace, a new classification method, called empirical classification, is developed for further signal dimension reduction in the classification step and has been shown to be more robust and outperform other simple classifiers. An overall accuracy of 98.34% for the classification of 689 real RS recording segments shows the promising performance of the presented method.

Adventitious sounds identification and extraction using temporal-spectral dominance-based features.

Respiratory sound (RS) signals carry significant information about the underlying functioning of the pulmonary system by the presence of adventitious sounds (ASs). Although many studies have addressed the problem of pathological RS classification, only a limited number of scientific works have focused on the analysis of the evolution of symptom-related signal components in joint time-frequency (TF) plane. This paper proposes a new signal identification and extraction method for various ASs based on instantaneous frequency (IF) analysis. The presented TF decomposition method produces a noise-resistant high definition TF representation of RS signals as compared to the conventional linear TF analysis methods, yet preserving the low computational complexity as compared to those quadratic TF analysis methods. The discarded phase information in conventional spectrogram has been adopted for the estimation of IF and group delay, and a temporal-spectral dominance spectrogram has subsequently been constructed by investigating the TF spreads of the computed time-corrected IF components. The proposed dominance measure enables the extraction of signal components correspond to ASs from noisy RS signal at high noise level. A new set of TF features has also been proposed to quantify the shapes of the obtained TF contours, and therefore strongly, enhances the identification of multicomponents signals such as polyphonic wheezes. An overall accuracy of 92.4±2.9% for the classification of real RS recordings shows the promising performance of the presented method.

Automatic wheeze detection using histograms of sample entropy.

In this paper, we propose a robust and automatic wheeze detection method using sample entropy (SampEn) histograms of the filtered narrow band respiratory sound signals. The sound signals are segmented first into their respective inspiration/expiration phases. Time-frequency distribution of each segment is then obtained using Gabor spectrogram. After the construction of SampEn plane, histograms of the selected frequency bins of the SampEn plane are calculated. The mean distortion of the histograms are used as discriminating features for segment-wise wheeze detection. Detection experiments are carried out irrespective of inspiration/expiration segments of the respiration sound signals recorded and preprocessed under different conditions, and the overall wheeze detection accuracy is 97.9% for high intensity wheezes during expirations and is up to 85.3% for low intensity wheezes occurring in inspirations.