Classification of Lung Sounds and Disease Prediction using Dense CNN Network

Respiratory illnesses are a main source of death in the world and exact lung sound identification is very significant for the conclusion and assessment of sickness. Be that as it may, this method is vulnerable to doctors and instrument limitations. As a result, the automated investigation and analysis of respiratory sounds has been a field of great research and exploration during the last decades. The classification of respiratory sounds has the potential to distinguish anomalies and diseases in the beginning phases of a respiratory dysfunction and hence improve the accuracy of decision making. In this paper, we explore the publically available respiratory sound database and deploy three different convolutional neural networks (CNN) and combine them to form a dense network to diagnose the respiratory disorders. The results demonstrate that this dense network classifies the sounds accurately and diagnoses the corresponding respiratory disorders associated with them.

Download Full-text

Automated classification of human lung sound signals using phase space representation of intrinsic mode function

10.21203/rs.3.rs-362731/v1 ◽

2021 ◽

Author(s):

Sibghatullah I. Khan ◽

Vikram Palodiya ◽

Lavanya Poluboyina

Keyword(s):

Phase Space ◽

Human Lung ◽

Feature Space ◽

Normal Lung ◽

Space Representation ◽

Support Vector ◽

Lung Sounds ◽

Lung Sound ◽

Phase Space Representation

Abstract Bronchiectasis and chronic obstructive pulmonary disease (COPD) are common human lung diseases. In general, the expert pulmonologistcarries preliminary screening and detection of these lung abnormalities by listening to the adventitious lung sounds. The present paper is an attempt towards the automatic detection of adventitious lung sounds ofBronchiectasis,COPD from normal lung sounds of healthy subjects. For classification of the lung sounds into a normaland adventitious category, we obtain features from phase space representation (PSR). At first, the empirical mode decomposition (EMD) is applied to lung sound signals to obtain intrinsic mode functions (IMFs). The IMFs are then further processed to construct two dimensional (2D) and three dimensional (3D) PSR. The feature space includes the 95% confidence ellipse area and interquartile range (IQR) of Euclidian distances computed from 2D and 3D PSRs, respectively. The process is carried out for the first four IMFs correspondings to normal and adventitious lung sound signals. The computed features depicta significant ability to discriminate the two categories of lung sound signals.To perform classification, we use the least square support vector machine with two kernels, namely, polynomial and radial basis function (RBF).Simulation outcomes on ICBHI 2017 lung sound dataset show the ability of the proposed method in effectively classifying normal and adventitious lung sound signals. LS-SVM is employing RBF kernel provides the highest classification accuracy of 97.67 % over feature space constituted by first, second, and fourth IMF.

Download Full-text

Towards the standardisation of lung sound nomenclature

European Respiratory Journal ◽

10.1183/13993003.01132-2015 ◽

2015 ◽

Vol 47 (3) ◽

pp. 724-732 ◽

Cited By ~ 35

Author(s):

Hans Pasterkamp ◽

Paul L.P. Brand ◽

Mark Everard ◽

Luis Garcia-Marcos ◽

Hasse Melbye ◽

...

Keyword(s):

Task Force ◽

Quality Criteria ◽

Supplementary Information ◽

Respiratory Sounds ◽

Lung Sounds ◽

Lung Sound ◽

European Respiratory Society ◽

Related Data ◽

Reference Collection ◽

E Learning

Auscultation of the lung remains an essential part of physical examination even though its limitations, particularly with regard to communicating subjective findings, are well recognised. The European Respiratory Society (ERS) Task Force on Respiratory Sounds was established to build a reference collection of audiovisual recordings of lung sounds that should aid in the standardisation of nomenclature. Five centres contributed recordings from paediatric and adult subjects. Based on pre-defined quality criteria, 20 of these recordings were selected to form the initial reference collection. All recordings were assessed by six observers and their agreement on classification, using currently recommended nomenclature, was noted for each case. Acoustical analysis was added as supplementary information. The audiovisual recordings and related data can be accessed online in the ERS e-learning resources. The Task Force also investigated the current nomenclature to describe lung sounds in 29 languages in 33 European countries. Recommendations for terminology in this report take into account the results from this survey.

Download Full-text

Convolutional neural networks based efficient approach for classification of lung diseases

Health Information Science and Systems ◽

10.1007/s13755-019-0091-3 ◽

2019 ◽

Vol 8 (1) ◽

Cited By ~ 8

Author(s):

Fatih Demir ◽

Abdulkadir Sengur ◽

Varun Bajaj

Keyword(s):

Neural Networks ◽

Convolutional Neural Networks ◽

Lung Diseases ◽

Support Vector ◽

Lung Sounds ◽

Lung Sound ◽

Frequency Method ◽

Time Frequency ◽

Sound Classification

AbstractTreatment of lung diseases, which are the third most common cause of death in the world, is of great importance in the medical field. Many studies using lung sounds recorded with stethoscope have been conducted in the literature in order to diagnose the lung diseases with artificial intelligence-compatible devices and to assist the experts in their diagnosis. In this paper, ICBHI 2017 database which includes different sample frequencies, noise and background sounds was used for the classification of lung sounds. The lung sound signals were initially converted to spectrogram images by using time–frequency method. The short time Fourier transform (STFT) method was considered as time–frequency transformation. Two deep learning based approaches were used for lung sound classification. In the first approach, a pre-trained deep convolutional neural networks (CNN) model was used for feature extraction and a support vector machine (SVM) classifier was used in classification of the lung sounds. In the second approach, the pre-trained deep CNN model was fine-tuned (transfer learning) via spectrogram images for lung sound classification. The accuracies of the proposed methods were tested by using the ten-fold cross validation. The accuracies for the first and second proposed methods were 65.5% and 63.09%, respectively. The obtained accuracies were then compared with some of the existing results and it was seen that obtained scores were better than the other results.

Download Full-text

Classification of Respiratory Sounds Including Normal and Crackle/Rhonchi Pathologies

2020 28th Signal Processing and Communications Applications Conference (SIU) ◽

10.1109/siu49456.2020.9302513 ◽

2020 ◽

Author(s):

Funda Cinyol ◽

Ugur Baysal ◽

Ethem Gelir ◽

Elif Babaoglu ◽

Sevinc Ulasli ◽

...

Keyword(s):

Respiratory Sounds

Download Full-text

Crackle Pitch Rises Progressively during Inspiration in Pneumonia, CHF, and IPF Patients

Pulmonary Medicine ◽

10.1155/2012/240160 ◽

2012 ◽

Vol 2012 ◽

pp. 1-4 ◽

Cited By ~ 1

Author(s):

Andrey Vyshedskiy ◽

Raymond Murphy

Keyword(s):

Heart Failure ◽

Lung Volume ◽

Interstitial Fibrosis ◽

Lung Sounds ◽

Lung Sound ◽

Lung Volumes ◽

Alternate Explanation ◽

40 Hz

Objective. It is generally accepted that crackles are due to sudden opening of airways and that larger airways produce crackles of lower pitch than smaller airways do. As larger airways are likely to open earlier in inspiration than smaller airways and the reverse is likely to be true in expiration, we studied crackle pitch as a function of crackle timing in inspiration and expiration. Our goal was to see if the measurement of crackle pitch was consistent with this theory.Methods. Patients with a significant number of crackles were examined using a multichannel lung sound analyzer. These patients included 34 with pneumonia, 38 with heart failure, and 28 with interstitial fibrosis.Results. Crackle pitch progressively increased during inspirations in 79% of all patients. In these patients crackle pitch increased by approximately 40 Hz from the early to midinspiration and by another 40 Hz from mid to late-inspiration. In 10% of patients, crackle pitch did not change and in 11% of patients crackle pitch decreased. During expiration crackle pitch progressively decreased in 72% of patients and did not change in 28% of patients.Conclusion. In the majority of patients, we observed progressive crackle pitch increase during inspiration and decrease during expiration. Increased crackle pitch at larger lung volumes is likely a result of recruitment of smaller diameter airways. An alternate explanation is that crackle pitch may be influenced by airway tension that increases at greater lung volume. In any case improved understanding of the mechanism of production of these common lung sounds may help improve our understanding of pathophysiology of these disorders.

Download Full-text

Classification of lung disorders based on lung sounds using deep learning

Journal of Physics Conference Series ◽

10.1088/1742-6596/1937/1/012036 ◽

2021 ◽

Vol 1937 (1) ◽

pp. 012036

Author(s):

Reshma P Joy ◽

Nishi Shahnaj Haider

Keyword(s):

Deep Learning ◽

Lung Sounds

Download Full-text

Reliability and validity of ICPC-2-R encoding by medical students

Revista Brasileira de Medicina de Família e Comunidade ◽

10.5712/rbmfc13(40)1655 ◽

2018 ◽

Vol 13 (40) ◽

pp. 1-6

Author(s):

Leonardo Ferreira Fontenelle ◽

Álvaro Damiani Zamprogno ◽

André Filipe Lucchi Rodrigues ◽

Lorena Camillato Sirtoli ◽

Natália Josiele Cerqueira Checon ◽

...

Keyword(s):

Primary Care ◽

Medical Students ◽

Reliability And Validity ◽

International Classification ◽

Free Text ◽

Reason For Encounter ◽

Community Physician ◽

In The Beginning ◽

Reasons For Encounter

Objective: To estimate how reliably and validly can medical students encode reasons for encounter and diagnoses using the International Classification of Primary Care, revised 2nd edition (ICPC-2-R). Methods: For every encounter they supervised during an entire semester, three family and community physician teachers entered the reasons for encounter and diagnoses in free text into a form. Two of four medical students and one teacher encoded each reason for encounter or diagnosis using the ICPC-2-R. In the beginning of the study, two three-hour workshops were held, until the teachers were confident the students were ready for the encoding. After all the reasons for encounter and the diagnoses had been independently encoded, the seven encoders resolved the definitive codes by consensus. We defined reliability as agreement between students and validity as their agreement with the definitive codes, and used Gwet’s AC1 to estimate this agreement. Results: After exclusion of encounters encoded before the last workshop, the sample consisted of 149 consecutive encounters, comprising 262 reasons for encounter and 226 diagnoses. The encoding had moderate to substantial reliability (AC1, 0.805; 95% CI, 0.767–0.843) and substantial validity (AC1, 0.864; 95% CI, 0.833–0.891). Conclusion: Medical students can encode reasons for encounter and diagnoses with the ICPC-2-R if they are adequately trained.

Download Full-text

Respiratory sound recorder – continued Innovation

Pediatrics (Suppl Consilium Medicum) ◽

10.26442/26586630.2021.2.201017 ◽

2021 ◽

pp. 146-151

Author(s):

Lada S. Starostina ◽

Natalia A. Geppe ◽

Vladimir S. Malyshev ◽

Saniia I. Valieva ◽

Irina L. Ginesina ◽

...

Keyword(s):

Respiratory Tract ◽

Computer Technology ◽

Respiratory Function ◽

Age Groups ◽

Patient Treatment ◽

Respiratory Sounds ◽

Early Age ◽

Respiratory Sound ◽

Respiratory Noises ◽

Inner Diameter

The study of external respiratory function (ERF) is important in the diagnosis of respiratory tract abnormalities in various diseases. In children, especially at an early age, there are many difficulties in conducting studies. In recent decades, due to the development of computer technology, there is great interest in the study of respiratory sounds, methods of their registration, processing and use in the assessment of the respiratory system in children and adults. Russian scientists have developed the method of respiratory airway sound investigation, which has proved its effectiveness, reliability and necessity of use in practice. Computer bronchophonography is based on the analysis of time and frequency characteristics of the spectrum of respiratory noises, arising from changes in the bronchial diameter due to increase in the stiffness of their walls or decrease in the inner diameter. Computed bronchophonography may be used for diagnostics of EFD disorders in patients of all age groups both in the in-patient and out-patient treatment.

Download Full-text