Lightweight Denoising Autoencoder Design for Noise Removal in Electrocardiography
This study proposes two denoising autoencoder models with discrete cosine transform and discrete wavelet transform, to remove electrode motion artifacts in noisy electrocardiography. Initially, the discrete cosine transform and discrete wavelet transform efficiently removed the high-frequency noise....
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2022-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/9889721/ |
_version_ | 1811265052520480768 |
---|---|
author | Ming-Hwa Sheu Yu-Syuan Jhang Yen-Ching Chang Szu-Ting Wang Chuan-Yu Chang Shin-Chi Lai |
author_facet | Ming-Hwa Sheu Yu-Syuan Jhang Yen-Ching Chang Szu-Ting Wang Chuan-Yu Chang Shin-Chi Lai |
author_sort | Ming-Hwa Sheu |
collection | DOAJ |
description | This study proposes two denoising autoencoder models with discrete cosine transform and discrete wavelet transform, to remove electrode motion artifacts in noisy electrocardiography. Initially, the discrete cosine transform and discrete wavelet transform efficiently removed the high-frequency noise. The six encoder layers then retain important electrocardiography features, whereas the six decoder layers reconstruct the clean electrocardiography. To improve the denoising performance, two network layers, the residual block and pixel adjustment, are added to the encoder and decoder layers to solve the vanishing gradient and improve subtle feature extraction. The proposed methods were applied to 66,000 real-recorded noisy electrocardiography fragments. The experimental result indicates that discrete wavelet transform based denoising autoencoder and discrete cosine transform based denoising autoencoder can improve the signal-to-noise ratio by 25.29 and 25.13 dB on average when the input signal-to-noise ratio is −6 dB. |
first_indexed | 2024-04-12T20:16:43Z |
format | Article |
id | doaj.art-147e388f9cd84341a68f13ae37bb1472 |
institution | Directory Open Access Journal |
issn | 2169-3536 |
language | English |
last_indexed | 2024-04-12T20:16:43Z |
publishDate | 2022-01-01 |
publisher | IEEE |
record_format | Article |
series | IEEE Access |
spelling | doaj.art-147e388f9cd84341a68f13ae37bb14722022-12-22T03:18:06ZengIEEEIEEE Access2169-35362022-01-0110981049811610.1109/ACCESS.2022.32066209889721Lightweight Denoising Autoencoder Design for Noise Removal in ElectrocardiographyMing-Hwa Sheu0Yu-Syuan Jhang1Yen-Ching Chang2Szu-Ting Wang3Chuan-Yu Chang4https://orcid.org/0000-0001-9476-8130Shin-Chi Lai5https://orcid.org/0000-0003-0011-3649Department of Electronics Engineering, National Yunlin University of Science & Technology, Douliu, TaiwanDepartment of Electronics Engineering, National Yunlin University of Science & Technology, Douliu, TaiwanDepartment of Electronics Engineering, National Yunlin University of Science & Technology, Douliu, TaiwanDoctor’s Program of Smart Industry Technology Research and Design, National Formosa University, Huwei, TaiwanDepartment of Computer Science and Information Engineering, National Yunlin University of Science & Technology, Yunlin, TaiwanDepartment of Automation Engineering, National Formosa University, Huwei, TaiwanThis study proposes two denoising autoencoder models with discrete cosine transform and discrete wavelet transform, to remove electrode motion artifacts in noisy electrocardiography. Initially, the discrete cosine transform and discrete wavelet transform efficiently removed the high-frequency noise. The six encoder layers then retain important electrocardiography features, whereas the six decoder layers reconstruct the clean electrocardiography. To improve the denoising performance, two network layers, the residual block and pixel adjustment, are added to the encoder and decoder layers to solve the vanishing gradient and improve subtle feature extraction. The proposed methods were applied to 66,000 real-recorded noisy electrocardiography fragments. The experimental result indicates that discrete wavelet transform based denoising autoencoder and discrete cosine transform based denoising autoencoder can improve the signal-to-noise ratio by 25.29 and 25.13 dB on average when the input signal-to-noise ratio is −6 dB.https://ieeexplore.ieee.org/document/9889721/Artificial neural networksbiomedical computingbiomedical signal processingdiscrete cosine transformsdeep learningdiscrete wavelet transforms |
spellingShingle | Ming-Hwa Sheu Yu-Syuan Jhang Yen-Ching Chang Szu-Ting Wang Chuan-Yu Chang Shin-Chi Lai Lightweight Denoising Autoencoder Design for Noise Removal in Electrocardiography IEEE Access Artificial neural networks biomedical computing biomedical signal processing discrete cosine transforms deep learning discrete wavelet transforms |
title | Lightweight Denoising Autoencoder Design for Noise Removal in Electrocardiography |
title_full | Lightweight Denoising Autoencoder Design for Noise Removal in Electrocardiography |
title_fullStr | Lightweight Denoising Autoencoder Design for Noise Removal in Electrocardiography |
title_full_unstemmed | Lightweight Denoising Autoencoder Design for Noise Removal in Electrocardiography |
title_short | Lightweight Denoising Autoencoder Design for Noise Removal in Electrocardiography |
title_sort | lightweight denoising autoencoder design for noise removal in electrocardiography |
topic | Artificial neural networks biomedical computing biomedical signal processing discrete cosine transforms deep learning discrete wavelet transforms |
url | https://ieeexplore.ieee.org/document/9889721/ |
work_keys_str_mv | AT minghwasheu lightweightdenoisingautoencoderdesignfornoiseremovalinelectrocardiography AT yusyuanjhang lightweightdenoisingautoencoderdesignfornoiseremovalinelectrocardiography AT yenchingchang lightweightdenoisingautoencoderdesignfornoiseremovalinelectrocardiography AT szutingwang lightweightdenoisingautoencoderdesignfornoiseremovalinelectrocardiography AT chuanyuchang lightweightdenoisingautoencoderdesignfornoiseremovalinelectrocardiography AT shinchilai lightweightdenoisingautoencoderdesignfornoiseremovalinelectrocardiography |