The Acoustic Dissection of Cough: Diving into Machine Listening-based COVID-19 Analysis and Detection

Abstract

Purpose: The coronavirus disease 2019 (COVID-19) has caused a crisis worldwide. Amounts of efforts have been made to prevent and control COVID-19's transmission, from early screenings to vaccinations and treatments. Recently, due to the spring up of many automatic disease recognition applications based on machine listening techniques, it would be fast and cheap to detect COVID-19 from recordings of cough, a key symptom of COVID-19. To date, knowledge on the acoustic characteristics of COVID-19 cough sounds is limited, but would be essential for structuring effective and robust machine learning models. The present study aims to explore acoustic features for distinguishing COVID-19 positive individuals from COVID-19 negative ones based on their cough sounds. Methods: With the theory of computational paralinguistics, we analyse the acoustic correlates of COVID-19 cough sounds based on the COMPARE feature set, i. e., a standardised set of 6,373 acoustic higher-level features. Furthermore, we train automatic COVID-19 detection models with machine learning methods and explore the latent features by evaluating the contribution of all features to the COVID-19 status predictions. Results: The experimental results demonstrate that a set of acoustic parameters of cough sounds, e. g., statistical functionals of the root mean square energy and Mel-frequency cepstral coefficients, are relevant for the differentiation between COVID-19 positive and COVID-19 negative cough samples. Our automatic COVID-19 detection model performs significantly above chance level, i. e., at an unweighted average recall (UAR) of 0.632, on a data set consisting of 1,411 cough samples (COVID-19 positive/negative: 210/1,201). Conclusions: Based on the acoustic correlates analysis on the COMPARE feature set and the feature analysis in the effective COVID-19 detection model, we find that the machine learning method to a certain extent relies on acoustic features showing higher effects in conventional group difference testing.

Is Speech the New Blood? Recent Progress in AI-Based Disease Detection From Audio in a Nutshell

“…A common approach is to derive a selection of features that leads to the best detection performance ( 37 ), or to identify those features that contribute most to the final model output. The knowledge about specific speech features that are most essential for the ML algorithm to differentiate between patients with a certain disease and healthy speakers, could allow the physician to draw conclusions about potential voice-physiological atypicalities associated with the investigated disease ( 38 ). Alternatively, sonification represents a recently emerging XAI approach, in which sound is generated to auditorily demonstrate salient facets of learning data or relevant acoustic features to allow human listeners to follow the reasoning of an ML algorithm ( 39 ).…”

Explainable COVID-19 detection using fractal dimension and vision transformer with Grad-CAM on cough sounds

“…Rahman et al [21] used CAMBRIDGE and QATARI datasets and obtained a 96.5% accuracy score. Ren et al [22] used the COUGHVID dataset and obtained a 0.632 Unweighted Average Recall (UAR) score. Andreu-Perez et al [25] used their own dataset with CNN and obtained 97.18% and 96.64% average accuracy scores for COVID-19 and healthy classes.…”

“…The combinations of the Cambridge and the Qatari datasets were used, and a 96.5% accuracy score was obtained using 5 fold cross validation test. Ren et al [22] used a set of 6,373 acoustic features based on COMPARE analysis to detect COVID-19 based on cough sound analysis. Authors used various machine learning methods to classify the acoustic features and explored that MFCC and bear-essential-acoustic information were quite efficient in COVID-19 detection.…”

A Hybrid Model Based on Deep Features and Ensemble Learning for the Diagnosis of COVID-19: DeepFeat-E

“…In addition, artificial intelligence systems can be successfully applied in a wide range of fields from lip-reading applications [8] to document language [9] and gesture recognition [10], epileptic seizures [11] and heart disease detection [12]. Similarly, artificial intelligence systems based on cough sounds [13,14] and especially chest images (X-Ray and CT scan) are widely used in COVID-19 diagnosis [15,16].…”