Heart sound and lung sound separation algorithms: a review

Nersisson, Ruban; Noel, Mathew Mithra

doi:10.1080/03091902.2016.1209589

Cited by 33 publications

(20 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Accordingly, the participant anesthesiologists were asked to rate the difficulty of operating the device during the auscultation on a scale of 0 to 10 (0: not difficult at all; 6: occasional difficulty during the test period; 10: extremely difficult during the test period). Furthermore, because the heart sound and lung sound are 2 major body sound signals from the chest region and often interfere with each other during auscultation, manufacturers provide unique algorithms for the separation of heart and lung sound ( 8 ). Therefore, lung sound quality in the questionnaire referred to the quality of the heart and lung sound separation during auscultation (0–10; 0: heart sound was comparable to or more obvious than lung sound with frequent heart sound interference during auscultation; 6: partial sound separation with occasional heart sound interference during auscultation; 10: complete sound separation without detectable heart sound during lung sound auscultation).…”

Section: Methodsmentioning

confidence: 99%

Use of Electronic Auscultation in Full Personal Protective Equipment to Detect Ventilation Status in Selective Lung Ventilation: A Randomized Controlled Trial

et al. 2022

View full text Add to dashboard Cite

Chest auscultation is the first procedure performed to detect endotracheal tube malpositioning but conventional stethoscopes do not conform to the personal protective equipment (PPE) protocol during the COVID-19 pandemic. This double-blinded randomized controlled trial evaluated the feasibility of using ear-contactless electronic stethoscope to identify endobronchial blocker established selective lung ventilation, simulating endobronchial intubation during thoracic surgery with full PPE. Conventional and electronic auscultation was performed without and with full PPE, respectively, of 50 patients with selective lung ventilation. The rates of correct ventilation status detection were 86 and 88% in the conventional and electronic auscultation groups (p = 1.00). Electronic auscultation revealed a positive predictive value of 87% (95% CI 77 to 93%), and a negative predictive value of 91% (95% CI 58 to 99%), comparable to the results for conventional auscultation. For detection of the true unilateral lung ventilation, the F1 score and the phi were 0.904 and 0.654, respectively for conventional auscultation; were 0.919 and 0.706, respectively for electronic auscultation. Furthermore, the user experience questionnaire revealed that the majority of participant anesthesiologists (90.5%) rated the audio quality of electronic lung sounds as comparable or superior to that of conventional acoustic lung sounds. In conclusion, electronic auscultation assessments of ventilation status as examined during thoracic surgery in full PPE were comparable in accuracy to corresponding conventional auscultation assessments made without PPE. Users reported satisfactory experience with the electronic stethoscope.

show abstract

Section: Methodsmentioning

confidence: 99%

Use of Electronic Auscultation in Full Personal Protective Equipment to Detect Ventilation Status in Selective Lung Ventilation: A Randomized Controlled Trial

et al. 2022

View full text Add to dashboard Cite

show abstract

“…For this purpose, 207 recordings were filtered with a 4 th -order Butterworth bandpass filter with passband frequencies 50-250 Hz and 200-1000 Hz, in order to separate heart and lung sounds, respectively. This is a commonly used approach to improve signal quality in commercial stethoscopes and clinical studies, with passband frequencies based on the main frequency bands reported in literature for neonatal heart and lung sounds [17].…”

Section: A Data Acquisition and Preprocessingmentioning

confidence: 99%

Real-Time Multi-Level Neonatal Heart and Lung Sound Quality Assessment for Telehealth Applications

Grooby,

Sitaula,

Fattahi

et al. 2021

Preprint

View full text Add to dashboard Cite

Through the usage of digital stethoscopes in combination with telehealth, chest sounds can be easily collected and transmitted for remote monitoring and diagnosis. Chest sounds contain important information about a newborn's cardiorespiratory health. However, low-quality recordings complicate the remote monitoring and diagnosis. In this study, a new method is proposed to objectively and automatically assess heart and lung signal quality on a 5-level scale in real-time, and to assess the effect of signal quality on vital sign estimation. For the evaluation, a total of 207 10 s long chest sounds were taken from 119 preterm and full-term babies. Thirty of the recordings from ten subjects were obtained with synchronous vital signs from the Neonatal Intensive Care Unit (NICU) based on electrocardiogram recordings. As reference, seven annotators independently assessed the signal quality. For automatic quality classification, 400 features were extracted from the chest sounds. After feature selection using minimum redundancy and maximum relevancy algorithm, class balancing, and hyper-parameter optimization, a variety of multi-class and ordinal classification and regression algorithms were trained. Then, heart rate and breathing rate were automatically estimated from the chest sounds using adapted preexisting methods. The results of subject-wise leave-one-out crossvalidation show that the best-performing models had a mean squared error (MSE) of 0.487 and 0.612, and balanced accuracy of 56.8% and 51.2% for heart and lung qualities, respectively. The best-performing models for real-time analysis (<200 ms) had MSE of 0.459 and 0.673, and balanced accuracy of 56.7% and 46.3%, respectively. Our experimental results underscore that increasing the signal quality leads to a reduction in vital sign error, with only high-quality recordings having mean absolute error of less than 5 beats per minute, as required for clinical usage.

show abstract

“…Our network design is inspired by the fundamental nature of heart sounds and their applicability in a real-world setting. The heart sound signal is quasi-periodic as it is generated at regular intervals by sequential opening and closing of heart valves as blood flows through heart chambers [22]. Therefore, we hypothesize that a recurrent module in the network will improve denoising performance compared to standalone convolutional models because of its capability to identify temporal relationships.…”

Section: Introductionmentioning

confidence: 99%

An End-to-end Deep Learning Framework for Real-Time Denoising of Heart Sounds for Effective Computer-Aided Auscultation in Unseen Noise

Ali¹,

Shuvo²,

Hasan³

2022

Preprint

View full text Add to dashboard Cite

<p>The heart sound signal captured via a digital stethoscope is often corrupted with environmental and physiological noise, altering its salient and critical properties. The problem is exacerbated in crowded low-resource hospital settings with high noise levels. In this study, inspired by the heart sound’s quasi-periodic nature, we present a novel deep encoder-decoder based LU-Net architecture to suppress ambient hospital and internal lung sound noises that corrupt the heart sounds. Training is done using a large benchmark PCG dataset mixed with physiological noise, i.e., breathing sounds. Two different noisy datasets were created for experimental evaluation by mixing unseen lung sounds and hospital ambiance noises with the clean heart sound recordings. We also use the inherently noisy portion of the PASCAL heart sound dataset for evaluation. Experimental results show that the proposed framework can effectively suppress background noise in both real-world and synthetically generated noisy heart sound recordings, improving the signal-to-noise ratio (SNR) level by 5.575 dB on an average using only 1.32 M parameters. The study reveals that the proposed model outperforms the current state-of-the-art U-Net model with an average SNR improvement of 5.613 dB and 5.537 dB in the presence of lung sound and hospital noise, respectively. LU-Net also outperformed the state-of-the-art Fully Convolutional Network (FCN) by 1.750 dB and 1.748 dB for lung sound and hospital noise conditions, respectively. In addition, the proposed denoising method model improves classification accuracy by 38.93% in the noisy portion of the PASCAL heart sound dataset. The proposed enhancement scheme can thus play a vital role in deploying automated cardiac screening systems in low-resource and underserved communities. </p>

show abstract

Heart sound and lung sound separation algorithms: a review

Cited by 33 publications

References 34 publications

Use of Electronic Auscultation in Full Personal Protective Equipment to Detect Ventilation Status in Selective Lung Ventilation: A Randomized Controlled Trial

Use of Electronic Auscultation in Full Personal Protective Equipment to Detect Ventilation Status in Selective Lung Ventilation: A Randomized Controlled Trial

Real-Time Multi-Level Neonatal Heart and Lung Sound Quality Assessment for Telehealth Applications

An End-to-end Deep Learning Framework for Real-Time Denoising of Heart Sounds for Effective Computer-Aided Auscultation in Unseen Noise

Contact Info

Product

Resources

About