Classification of Aortic Stenosis Using Time–Frequency Features From Chest Cardio-Mechanical Signals

Yang, Chenxi; Aranoff, Nicole; Tavassolian, Negar

doi:10.1109/tbme.2019.2942741

Cited by 39 publications

(49 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared to the ANOVA results from our previous study 22 , the frequency and statistical distributions show similar observations. The significant CWT features from 22 are below 10 Hz.…”

Section: Feature Analysis and Selection Results For Conventional ML Msupporting

confidence: 78%

“…The percentage of maximum-based features are higher in this study, with values of more than 80%. Since 22 only uses single-axis SCG and GCG signals, there is no fair comparison between source distributions. The differences in frequency and statistical distributions might be caused by the removal of age bias and the inclusion of multi-dimensional SCG and GCG recordings.…”

Section: Feature Analysis and Selection Results For Conventional ML Mmentioning

confidence: 99%

“…There are two main differences in this step between our current and previous frameworks 22 . The first difference is that more dimensions from SCG and GCG recordings have been included in feature generation.…”

Section: Time-frequency Conversion Of Temporal Signals Each Of Thosementioning

confidence: 96%

“…Secondly, we did not use the empirical mode decomposition in the current study. The reason is that much fewer features were selected from EMD compared to CWT in 22 . The ranks of the EMD-based features were also relatively low in ANOVA tests.…”

Section: Time-frequency Conversion Of Temporal Signals Each Of Thosementioning

confidence: 99%

“…A root mean square (RMS) filter with a step size of 500 ms was used for segmentation, and a threshold of 1.5 times the median of the RMS value was implemented to select low-noise segments. The detail for this signal processing and filtering is included in our previous study in 22 . As a result, a total of 910 segments were created for all axes and modalities.…”

mentioning

confidence: 99%

See 4 more Smart Citations

Classification of aortic stenosis using conventional machine learning and deep learning methods based on multi-dimensional cardio-mechanical signals

Yang

Ojha

Aranoff

et al. 2020

Sci Rep

Self Cite

View full text Add to dashboard Cite

This paper introduces a study on the classification of aortic stenosis (AS) based on cardio-mechanical signals collected using non-invasive wearable inertial sensors. Measurements were taken from 21 AS patients and 13 non-AS subjects. A feature analysis framework utilizing Elastic Net was implemented to reduce the features generated by continuous wavelet transform (CWT). Performance comparisons were conducted among several machine learning (ML) algorithms, including decision tree, random forest, multi-layer perceptron neural network, and extreme gradient boosting. In addition, a two-dimensional convolutional neural network (2D-CNN) was developed using the CWT coefficients as images. The 2D-CNN was made with a custom-built architecture and a CNN based on Mobile Net via transfer learning. After the reduction of features by 95.47%, the results obtained report 0.87 on accuracy by decision tree, 0.96 by random forest, 0.91 by simple neural network, and 0.95 by XGBoost. Via the 2D-CNN framework, the transfer learning of Mobile Net shows an accuracy of 0.91, while the custom-constructed classifier reveals an accuracy of 0.89. Our results validate the effectiveness of the feature selection and classification framework. They also show a promising potential for the implementation of deep learning tools on the classification of AS.

show abstract

“…Compared to the ANOVA results from our previous study 22 , the frequency and statistical distributions show similar observations. The significant CWT features from 22 are below 10 Hz.…”

Section: Feature Analysis and Selection Results For Conventional ML Msupporting

confidence: 78%

Section: Feature Analysis and Selection Results For Conventional ML Mmentioning

confidence: 99%

Section: Time-frequency Conversion Of Temporal Signals Each Of Thosementioning

confidence: 96%

Section: Time-frequency Conversion Of Temporal Signals Each Of Thosementioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations

Classification of aortic stenosis using conventional machine learning and deep learning methods based on multi-dimensional cardio-mechanical signals

Yang

Ojha

Aranoff

et al. 2020

Sci Rep

Self Cite

View full text Add to dashboard Cite

show abstract

A Scalable Laser‐Centric Fabrication of an Epidermal Cardiopulmonary Patch

Rachim

Lee

Kim

et al. 2022

Adv Materials Technologies

View full text Add to dashboard Cite

Recent advances in scalable fabrication methods based on printing technologies have improved the yield and lowered the cost of manufacturing epidermal sensors. However, modern technologies still require expensive multi‐bio‐ink raw materials. A laser‐centric fabrication method that realizes a cost‐effective, scalable, and streamlined fabrication process with easily used material and equipment is reported. The fabricated epidermal patch can quickly respond to the demands of personalized models. Here, the epidermal sensor patch is applied to continuous monitoring of the human cardiopulmonary system. The sensor performs conventional bio‐signal monitoring but is also extendible to other monitoring forms. The patch provides enhanced functions through its gas‐permeable and skin‐adhesive microporous layer and its stretchable, conformal, and biocompatible multimodal sensing layer. The utility of the proposed sensor patch in clinical diagnosis in a mobile healthcare environment is demonstrated in multiple bio‐signal‐morphology analyses.

show abstract