An Embedded Classifier of Lung Sounds based on the Wavelet Packet Transform and ANN

BackgroundAutomatic detection or classification of adventitious sounds is useful to assist physicians in diagnosing or monitoring diseases such as asthma, Chronic Obstructive Pulmonary Disease (COPD), and pneumonia. While computerised respiratory sound analysis, specifically for the detection or classification of adventitious sounds, has recently been the focus of an increasing number of studies, a standardised approach and comparison has not been well established.ObjectiveTo provide a review of existing algorithms for the detection or classification of adventitious respiratory sounds. This systematic review provides a complete summary of methods used in the literature to give a baseline for future works.Data sourcesA systematic review of English articles published between 1938 and 2016, searched using the Scopus (1938-2016) and IEEExplore (1984-2016) databases. Additional articles were further obtained by references listed in the articles found. Search terms included adventitious sound detection, adventitious sound classification, abnormal respiratory sound detection, abnormal respiratory sound classification, wheeze detection, wheeze classification, crackle detection, crackle classification, rhonchi detection, rhonchi classification, stridor detection, stridor classification, pleural rub detection, pleural rub classification, squawk detection, and squawk classification.Study selectionOnly articles were included that focused on adventitious sound detection or classification, based on respiratory sounds, with performance reported and sufficient information provided to be approximately repeated.Data extractionInvestigators extracted data about the adventitious sound type analysed, approach and level of analysis, instrumentation or data source, location of sensor, amount of data obtained, data management, features, methods, and performance achieved.Data synthesisA total of 77 reports from the literature were included in this review. 55 (71.43%) of the studies focused on wheeze, 40 (51.95%) on crackle, 9 (11.69%) on stridor, 9 (11.69%) on rhonchi, and 18 (23.38%) on other sounds such as pleural rub, squawk, as well as the pathology. Instrumentation used to collect data included microphones, stethoscopes, and accelerometers. Several references obtained data from online repositories or book audio CD companions. Detection or classification methods used varied from empirically determined thresholds to more complex machine learning techniques. Performance reported in the surveyed works were converted to accuracy measures for data synthesis.LimitationsDirect comparison of the performance of surveyed works cannot be performed as the input data used by each was different. A standard validation method has not been established, resulting in different works using different methods and performance measure definitions.ConclusionA review of the literature was performed to summarise different analysis approaches, features, and methods used for the analysis. The performance of recent studies showed a high agreement with conventiona...

show abstract

“…An ANN was used in [ 71 ] to perform event classification of respiratory sounds containing wheezes and crackles. Data was obtained from [ 129 ].…”

Section: Resultsmentioning

confidence: 99%

Automatic adventitious respiratory sound analysis: A systematic review

2017

View full text Add to dashboard Cite

show abstract

“…Embora muitas aplicações exijam um tempo extremamente reduzido para processar os dados, são relatados na literatura trabalhos que empregam com sucesso sistemas embarcados para realizar tarefas de classificação em tempo real (Shi et al, 2009;Rúa et al, 2012;Tocchetto et al, 2014). Contudo, ainda resta estabelecer uma conexão mais consistente entre classificadores gerados off-board e sua utilização em sistemas embarcados, dada a pouca variedade de modelos de classificação empregados e a pouca disponibilidade de ferramentas que realizem adequadamente essa interface.…”

Section: Introductionunclassified

“…O foco da nossa ferramentaé gerar implementações de classificadores especificamente para esse tipo de sistema, com programação bare metal. Inclusive, as escolhas dos modelos de classificação suportados ratificam esse propósito, dado que são modelos simples, como regressão logística eárvore de decisão (J48), ou que já foram empregados com sucesso em aplicações embarcadas, como Multilayer Perceptron (MLP) e Support Vector Machine (SVM) (Shi et al, 2009;Rúa et al, 2012;Sampaio et al, 2017;Farooq et al, 2010;Tocchetto et al, 2014). A intençãoé que nossa ferramenta seja genérica, suportando diferentes plataformas embarcadas, e que permita melhorar o desempenho dos classificadores através do uso de representação de pontofixo em vez de ponto flutuante.…”

Section: Introductionunclassified

Uma Ferramenta de Suporte ao Uso de Classificadores em Sistemas Embarcados

Silva

Souza

Batista

2019

Anais Do 14º Simpósio Brasileiro De Automação Inteligente

View full text Add to dashboard Cite

Computer systems are employed to solve existing problems in different contexts in society, such as health, industry, safety, transportation, and others. Many of these solutions are composed of sensors, embedded systems, and wireless communication. Moreover, researchers and practitioners have used machine learning models in embedded systems to perform complex tasks that can not be directly solved by traditional algorithms. In this work, we propose a software tool that converts classification models, trained with WEKA, to produce suitable classifiers to run in low-cost embedded systems with the option to use fixed-point representation. From an experimental evaluation with different embedded platforms, datasets and classifiers, we show consistent results of accuracy and a reduction of up to 92.9% in the classification time when using fixed-point representation. Resumo: Sistemas computacionais são empregados para solucionar problemas existentes em diferentes contextos na sociedade, como saúde, indústria, segurança, transporte, entre outros. Muitas dessas soluções são compostas por sensores, sistemas embarcados e comunicação sem fio. Além disso, academia e indústria têm considerado a utilização de modelos de Aprendizado de Máquina em sistemas embarcados para realizar tarefas complexas que não podem ser diretamente solucionadas por algoritmos tradicionais. Neste trabalho,é proposta uma ferramenta de software que converte modelos de classificação, treinados com a ferramenta WEKA, para produzir classificadores que executarão em sistemas embarcados de baixo custo, com a opção de utilizar representação de ponto-fixo. A partir de uma avaliação experimental com diferentes plataformas embarcadas, conjuntos de dados e classificadores, são demonstrados resultados consistentes de acurácia e redução de até 92, 9% no tempo de classificação ao utilizar representação de ponto-fixo.

show abstract

“…However, most ML algorithms output classifiers that are highly efficient, including the ones supported in our tool: Logistic Regression, Decision Tree, Multilayer Perceptron (MLP), and Support Vector Machine (SVM). The choice of these approaches is also based on the literature that reports the successful application of these methods with embedded hardware (FAROOQ et al, 2010;RÚA et al, 2012;TOCCHETTO et al, 2014;SAMPAIO et al, 2017).…”

Section: Proposalmentioning

confidence: 99%

“…A system for real-time monitoring and classification of lung signals, proposed by Tocchetto et al (2014), aims to develop a portable device for clinical support. In this proposal, the device captures lung sounds through a microphone placed in the patient's thorax region, processes the audio signals, and sends the information to a computer server that is available to the attending doctor.…”

Section: Lung Signals Classificationmentioning

confidence: 99%

Evaluating classification models for resource-constrained hardware

Silva¹

View full text Add to dashboard Cite

An Embedded Classifier of Lung Sounds based on the Wavelet Packet Transform and ANN

Cited by 18 publications

References 13 publications

Automatic adventitious respiratory sound analysis: A systematic review

Automatic adventitious respiratory sound analysis: A systematic review

Uma Ferramenta de Suporte ao Uso de Classificadores em Sistemas Embarcados

Evaluating classification models for resource-constrained hardware

Contact Info

Product

Resources

About