Machine learning for detecting COVID-19 from cough sounds: An ensemble-based MCDM method

Chowdhury, Nihad Karim; Kabir, Muhammad Ashad; Rahman, Md. Muhtadir; Islam, Sheikh Mohammed Shariful

doi:10.1016/j.compbiomed.2022.105405

Cited by 70 publications

(60 citation statements)

References 47 publications

(64 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The participants record their cough and breath sounds using devices, such as mobile phones and computers, and then upload them to a sound database. Several sound datasets have been developed, including Cambridge [116], NoCoCoDa [117], Coswara [118], COUGH-VID [119], Virufy [120], and they have been widely used to validate the feasibility of using sound for COVID-19 detection [121,122]. The collected data need to be analyzed and classified to determine whether someone has been infected with COVID-19 virus (Fig.…”

Section: Sound-based Diagnosis Of Covid-19mentioning

confidence: 99%

Visual Detection of COVID-19 from Materials Aspect

Wang

Meng

et al. 2022

Adv. Fiber Mater.

View full text Add to dashboard Cite

In the recent COVID-19 pandemic, World Health Organization emphasized that early detection is an effective strategy to reduce the spread of SARS-CoV-2 viruses. Several diagnostic methods, such as reverse transcription-polymerase chain reaction (RT-PCR) and lateral flow immunoassay (LFIA), have been applied based on the mechanism of specific recognition and binding of the probes to viruses or viral antigens. Although the remarkable progress, these methods still suffer from inadequate cellular materials or errors in the detection and sampling procedure of nasopharyngeal/oropharyngeal swab collection. Therefore, developing accurate, ultrafast, and visualized detection calls for more advanced materials and technology urgently to fight against the epidemic. In this review, we first summarize the current methodologies for SARS-CoV-2 diagnosis. Then, recent representative examples are introduced based on various output signals (e.g., colorimetric, fluorometric, electronic, acoustic). Finally, we discuss the limitations of the methods and provide our perspectives on priorities for future test development. Graphical Abstract

show abstract

Section: Sound-based Diagnosis Of Covid-19mentioning

confidence: 99%

Visual Detection of COVID-19 from Materials Aspect

Wang

Meng

et al. 2022

Adv. Fiber Mater.

View full text Add to dashboard Cite

show abstract

“…In [5] , the author provided a comprehensive study of the work done so for COVID-19 detection using deep learning techniques and frameworks, especially defining various components of the works and comparing using various existing pre-trained models and the datasets used in different studies. According to this study, the Resnet-50 model is the most used learning model for COVID-19 detection [5] , [13] , [14] , [15] . In this work also the mentioned accuracy is not enough to predict the COVID patient.…”

Section: Literature Workmentioning

confidence: 99%

“…The traditional clinical procedural method is used for the early diagnosis of infected people [1] , [2] , [3] . However, it takes more time to diagnose COVID-19 patients [5] . The pathological testing methods also are a very time-consuming process for the diagnosis of patients.…”

Section: Introductionmentioning

confidence: 99%

“…Among the proposed framework-based solutions, deep multimodal learning-based and Artificial Intelligence (AI) learning models are getting proliferation due to broad applications. It plays an essential role in fighting the COVID-19 pandemic based on several datasets such as chest-X-rays and cough sound-based datasets for better analysis [5] , [13] , [14] , [15] . No literature has yet been published to the best of our knowledge based on multimodal-based early diagnosis of the spread of COVID-19.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Ensemble multimodal deep learning for early diagnosis and accurate classification of COVID-19

Kumar

Gupta

Kumar

et al. 2022

Computers and Electrical Engineering

View full text Add to dashboard Cite

“…To this goal, four pretrained CNN models were utilized, and the suggested technique improved the area under the receiver operating characteristic curve by 3.57 percent over the baseline of DiCOVA Track-1 validation. Chowdhury et al [15] suggested an ensemble-based multi-criteria decision-making strategy for selecting the best COVID-19 cough classification machine learning methodology. To test their strategy, the authors employed four cough datasets: Cambridge, Coswara, Virufy, and NoCoCoDa.…”

Section: Introductionmentioning

confidence: 99%