A Wireless Electronic Stethoscope to Classify Children Heart Sound Abnormalities

Islam, Md. Riadul; Hassan, Md. Mahadi; Raihan, M.; Datto, Sabuz Kumar; Shahriar, Abdullah; More, Arun

doi:10.1109/iccit48885.2019.9038406

Cited by 7 publications

(13 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Significant research progress has been made in developing decision support systems based on various machine learning and deep learning algorithms for heart disease diagnosis which poses a serious health concern for the general population. Electrocardiograms (ECG or EKG) [29], [22], [6], [23], [30], [8], Photoplethysmograph [27], [36], [9], [26], [13], and auscultations [34], [14], [31], [32] modalities have been used for automated decision making. Owing to large body of work on deep learning based heart disease classification, in this section we limit the scope of our discussion to the most relevant works to our approach of using UMAP and research on the Mendeley Children heart Sound dataset.…”

Section: Related Workmentioning

confidence: 99%

“…On the Mendeley Children heart Sound dataset, authors Islam et. al [14] in their work converted the signals into MFCC, engineered features and presented comparisons of various SVM kernels. Authors Rani et.…”

Section: Related Workmentioning

confidence: 99%

“…We select the Mendeley Children heart sound dataset of normal and abnormal pediatric heart sounds from the rural areas of Bangladesh [14] collected from 60 subjects using their developed wireless electronic stethoscope. The collected dataset of 1657 samples, is preprocessed by re-sampling to 44100 Hz, normalized and denoised using Discrete Wavelet Transform.…”

Section: Datasetmentioning

confidence: 99%

“…Authors Islam et. al in their work proposed the dataset [14] and achieved 94.12% test accuracy, 88.89% specificity, and 100% sensitivity values using RBF SVM kernel on engineered MFCC features. In [31] 93.76% test accuracy was achieved using Convolution Neural Network using MFCC.…”

Section: B Comparison On Same Datasetmentioning

confidence: 99%

“…One of the significant challenges that hinders the availability of large quantities of data required for training data hungry deep learning models is the cost of data collection, limiting the accessibility of deep learning based decision support systems to general public specially in developing regions of the world [14]. This has motivated the development of low-cost diagnostic devices such as wireless phone stethoscope [14], wireless OCT devices [25]. Data collection using these lowcost devices can be noisy and have out-of-distribution samples collectively termed as outliers.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Multiview Outlier Filtered Pediatric Heart Sound Classification

Dakshit

2024

ijacsa

View full text Add to dashboard Cite

The advancements in deep learning has generated a large-scale interest in development of black-box models for various use cases in different domains such as healthcare, in both at-home and critical setting for diagnosis and monitoring of various health conditions. The use of audio signals as a view for diagnosis is nascent and the success of deep learning models in ingesting multimedia data provides an opportunity for use as a diagnostic medium. For the widespread use of these decision support systems, it is prudent to develop high performing systems which require large quantities of data for training and low-cost method of data collection making it more accessible for developing regions of the world and general population. Data collected from low-cost collection especially wireless devices are prone to outliers and anomalies. The presence of outliers skews the hypothesis space of the model and leads to model drift on deployment. In this paper, we propose a multiview pipeline through interpretable outlier filtering on the small Mendeley Children Heart Sound dataset collected using wireless low-cost digital stethoscope. Our proposed pipeline explores and provides dimensionally reduced interpertable visualizations for functional understanding of the effect of various outlier filtering methods on deep learning model hypothesis space and fusion strategies for multiple views of heart sound data namely raw time-series signal and Mel Frequency Cepstrum Coefficients achieving 98.19% state-of-the-art testing accuracy.

show abstract