Signal analysis of medical acoustic sounds with applications to chest medicine

doi:10.1016/j.jfranklin.2006.08.003

Journal of the Franklin Institute

Volume 344, Issues 3–4, May–July 2007, Pages 230-242

https://doi.org/10.1016/j.jfranklin.2006.08.003 Get rights and content

Abstract

The stethoscope is widely used for listening to heart sounds (cardiology), lung sounds (chest medicine) and digestive, bowel sounds, etc. An obvious development from this is to seek to automate the process by which a physician listens to and interprets the sounds. In parallel with signal processing developments, medical usage of diagnostic instrumentation is changing. The study described in this paper aimed to prove the feasibility of an automatic stethoscope for particular screening and monitoring algorithms, for which we envisage genuine clinical applicability. The first algorithm is to enable non-specialists to screen for pulmonary fibrosis. The algorithm was found to distinguish effectively between signals representing patients with chronic obstructive pulmonary disease (COPD), those with pulmonary fibrosis and normal individuals. A clear demarcation was observed between signals from patients in different diagnostic categories, with orders of magnitude differences between the groups. The second application is the monitoring of asthmatic patients. The proportion of wheeze within a given time recording is an objective indicator of the patient's condition, which can be used to monitor the person's progress during treatment. The proposed usage of each algorithm is demonstrated through a conceptual device, in which the ability to objectively quantify the patient's condition represents a powerful opportunity to improve clinical outcomes.

Introduction

The idea of listening to the sounds a body makes and deriving medical information from it is well established in medical practice and literature. The stethoscope was invented by the French physician Laënnec in 1816, originally for listening to respiratory sounds [1]. It is now also widely used for listening to heart sounds (cardiology) and digestive, bowel sounds, etc.

An obvious development from this has been to seek to automate the process by which a physician listens to and interprets the sounds. Interestingly, in the case of cardiology, where possibly the most significant medical impact could be made, electrical signal measurements (electrocardiography, ECG) offer such a superior signal to noise ratio that these have been developed instead and the use of stethoscopes in cardiology has declined in recent years. No such replacement exists as yet in chest medicine, or indeed in many of the other areas of medicine in which the stethoscope has a place. Furthermore, the stethoscope remains attractive as a ‘low tech’, relatively low cost instrument, which can be used, at least at a basic level, with relatively little training. However, the development of automatic lung sound analysis capabilities has been remarkably elusive, despite significant research activity in the area.

This paper reviews work in the analysis of heart, lung and other bodily sounds, presents the authors’ own recent advances in lung sounds analysis and discusses these in the context of challenges to apply these advances. As signal processing researchers, can we make a useful contribution in this area?

Section snippets

Automatic recognition of medical acoustic sounds

Categorisation of medical acoustic sounds, through the subjective ear of the physician, followed naturally from Laënnec's invention. Descriptions of the characteristic ‘adventitious’ (or abnormal) sounds are described in [2], together with an explanation of the changes to the lung physiology which cause the sounds. A more recent treatment of the subject is given in [3], with a particular view to reviewing lung sound recording sensors and proposing new directions. A further update on the subject

Lung sounds analysis

Clinically qualified staff collected data using a Meditron stethoscope [21], without selecting the built in filters, and converted it into PC wave files for analysis. The stethoscope is shown in Fig. 1. This process was done either by direct transfer to a laptop PC via the Meditron proprietary recording software or via a commercial minidisk recorder and then manual transfer using the Audacity freeware programme [22]. For each recording, a physician pre-diagnosed one of four different lung

Practical applications

In parallel with signal processing developments, medical usage of diagnostic instrumentation is changing. Within the UK, the original National Health Service model of a highly centralised, highly skilled and knowledgeable key decision maker (the physician) is being changed at policy level to a more decentralised, ‘patient centred’ and nurse-led structure [24]. Similar policies are evolving in other countries. In this new structure, other professionals, particularly nurses, will be more involved

Conclusion

The evolution of the stethoscope into an electronic transducer is paralleled by a development of hardware and software, making the automatic stethoscope a reality. A number of researchers, including ourselves, have demonstrated that successful applications of the automatic stethoscope in both cardiology and chest medicine are possible. We have presented results in which pulmonary fibrosis can be screened and asthmatic wheeze monitored, through two different algorithms. The process by which

Acknowledgements

We are grateful to Dr. S.B. Pearson, Leeds General Infirmary, Leeds for assistance in providing data for this study and for clinical guidance. Further data was kindly provided by Dr B.Johansen, Rijkshospital, Oslo, Norway. Electronic stethoscopes were supplied to the project by Meditron, Norway and Welch Allyn plc.

References (26)

P. Forgacs
Crackles and wheezes
Lancet
(1967)
D. Chen et al.
Estimation of the pulmonary artery pressure by spectral analysis of the second heart sound
Am. J. Cardiol.
(1996)
D. Kim et al.
Assessment of the severity of aortic stenosis through time–frequency analysis of murmur
Chest
(2003)
R.T.H. Laënnec
De l’auscultation médiate ou traité du diagnostic de maladies des poumons et du coeur, fondé principalement sur ce nouveau moyen d’exploration
(1819)
H. Pasterkamp et al.
Respiratory sounds—advances beyond the stethoscope
Am. J. Respir. Crit. Care Med.
(1997)
A. Sorvijärvi et al.
Computerised respiratory sound analysis (CORSA): recommended standards for terms and techniques
Eur. Resp. Rev.
(2000)
R.S. Khandpur
Biomedical Instrumentation: Technology and Applications
(2005)
R.M. Rangayyan et al.
Phonocardiogram signal analysis: a review
Crit. Rev. Biomed. Eng.
(1987)
A. Voss et al.
Diagnosing aortic valve stenosis by parameter extraction of heart sound signals
Ann. Biomed. Eng.
(2005)
M. El-Segaier et al.
Computer-based detection and analysis of heart sound and murmur
Ann. Biomed. Eng.
(2005)

J.C. Goswami et al.

Fundamentals of Wavelets: Theory, Algorithms and Applications

(1999)

L.J. Hadjileontiadis

Wavelet-based enhancement of lung and bowel sounds using fractal dimension thresholding—Part I: Methodology

IEEE Trans. Biomed. Eng.

(2005)

L.J. Hadjileontiadis

Wavelet-based enhancement of lung and bowel sounds using fractal dimension thresholding—Part II: Application results

IEEE Trans. Biomed. Eng.

(2005)

Cited by (46)

A novel feature extraction technique for pulmonary sound analysis based on EMD
2018, Computer Methods and Programs in Biomedicine
Citation Excerpt :
Also, there may be misinterpretation regarding the pulmonary dysfunctions due to lack of clarity of the signals’ information and also lack of knowledge in correlating sound signals characteristics and pathological signatures. The alternative to auscultation technique are based on image processing methods such as chest X-ray, roentgenogram and computerized tomography (CT) scan [4]. These tools produce an image of the region within the respiratory system that helps the physicians to visualize and make an interpretation about the diseased conditions.
Background and objective: The stethoscope based auscultation technique is a primary diagnostic tool for chest sound analysis. However, the performance of this method is limited due to its dependency on physicians experience, knowledge and also clarity of the signal. To overcome this problem we need an automated computer-aided diagnostic system that will be competent in noisy environment. In this paper, a novel feature extraction technique is introduced for discriminating various pulmonary dysfunctions in an automated way based on pattern recognition algorithms.
Method: In this work, the disease correlated relevant characteristics of lung sounds signals are identified in terms of statistical distribution parameters: mean, variance, skewness, and kurtosis. These features are extracted from selective morphological components of the mapped signal in the empirical mode decomposition domain. The feature set is fed to the classifier model to differentiate their corresponding classes.
Results: The significance of features developed are validated by conducting several experiments using supervised and unsupervised classifiers. Furthermore, the discriminating power of the proposed features is compared with three types of baseline features. The experimental result is evaluated by statistical analysis and also validated with physicians inference.
Conclusions: It is found that the proposed features extraction technique is superior to the baseline methods in terms of classification accuracy, sensitivity and specificity. The developed method gives better results compared to baseline methods in any circumstance. The proposed method gives a higher accuracy of 94.16, sensitivity of 100 and specificity of 93.75 for an artificial neural network classifier.
Construction of Ergodic GMM-HMMs for Classification between Healthy Individuals and Patients Suffering from Pulmonary Disease
2024, IEICE Transactions on Information and Systems
Acoustic HMMs to Detect Abnormal Respiration with Limited Training Data
2023, IEICE Transactions on Information and Systems
Patient Detection Using Inhalations or Exhalations in a Series of Respirations
2023, 6th International Seminar on Research of Information Technology and Intelligent Systems, ISRITI 2023 - Proceeding
Examination of Training Data Expansion for Detection of Abnormal Respiration and Patients
2023, Proceedings of the 2023 IEEE International Conference on Industry 4.0, Artificial Intelligence, and Communications Technology, IAICT 2023
Design of Lung Sound Acquisition and Analysis Device
2022, ACM International Conference Proceeding Series

View all citing articles on Scopus

View full text

Signal analysis of medical acoustic sounds with applications to chest medicine

Abstract

Introduction

Section snippets

Automatic recognition of medical acoustic sounds

Lung sounds analysis

Practical applications

Conclusion

Acknowledgements

Lancet

Am. J. Cardiol.

Chest

De l’auscultation médiate ou traité du diagnostic de maladies des poumons et du coeur, fondé principalement sur ce nouveau moyen d’exploration

Respiratory sounds—advances beyond the stethoscope

Am. J. Respir. Crit. Care Med.

Computerised respiratory sound analysis (CORSA): recommended standards for terms and techniques

Eur. Resp. Rev.

Biomedical Instrumentation: Technology and Applications

Phonocardiogram signal analysis: a review

Crit. Rev. Biomed. Eng.

Diagnosing aortic valve stenosis by parameter extraction of heart sound signals

Ann. Biomed. Eng.

Computer-based detection and analysis of heart sound and murmur

Ann. Biomed. Eng.

Fundamentals of Wavelets: Theory, Algorithms and Applications

Wavelet-based enhancement of lung and bowel sounds using fractal dimension thresholding—Part I: Methodology

IEEE Trans. Biomed. Eng.

Wavelet-based enhancement of lung and bowel sounds using fractal dimension thresholding—Part II: Application results

IEEE Trans. Biomed. Eng.