Cargando…

Electrical Impedance Tomography Based on Grey Wolf Optimized Radial Basis Function Neural Network

Electrical impedance tomography (EIT) is a non-invasive, radiation-free imaging technique with a lot of promise in clinical monitoring. However, since EIT image reconstruction is a non-linear, pathological, and ill-posed issue, the quality of the reconstructed images needs constant improvement. To i...

Descripción completa

Detalles Bibliográficos
Autores principales: Wang, Guanghua, Feng, Di, Tang, Wenlai
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9322610/
https://www.ncbi.nlm.nih.gov/pubmed/35888936
http://dx.doi.org/10.3390/mi13071120
_version_ 1784756347251720192
author Wang, Guanghua
Feng, Di
Tang, Wenlai
author_facet Wang, Guanghua
Feng, Di
Tang, Wenlai
author_sort Wang, Guanghua
collection PubMed
description Electrical impedance tomography (EIT) is a non-invasive, radiation-free imaging technique with a lot of promise in clinical monitoring. However, since EIT image reconstruction is a non-linear, pathological, and ill-posed issue, the quality of the reconstructed images needs constant improvement. To increase image reconstruction accuracy, a grey wolf optimized radial basis function neural network (GWO-RBFNN) is proposed in this paper. The grey wolf algorithm is used to optimize the weights in the radial base neural network, determine the mapping between the weights and the initial position of the grey wolf, and calculate the optimal position of the grey wolf to find the optimal solution for the weights, thus improving the image resolution of EIT imaging. COMSOL and MATLAB were used to numerically simulate the EIT system with 16 electrodes, producing 1700 simulation samples. The standard Landweber, RBFNN, and GWO-RBFNN approaches were used to train the sets separately. The obtained image correlation coefficient (ICC) of the test set after training with GWO-RBFNN is 0.9551. After adding 30, 40, and 50 dB of Gaussian white noise to the test set, the attained ICCs with GWO-RBFNN are 0.8966, 0.9197, and 0.9319, respectively. The findings reveal that the proposed GWO-RBFNN approach outperforms the existing methods when it comes to image reconstruction.
format Online
Article
Text
id pubmed-9322610
institution National Center for Biotechnology Information
language English
publishDate 2022
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-93226102022-07-27 Electrical Impedance Tomography Based on Grey Wolf Optimized Radial Basis Function Neural Network Wang, Guanghua Feng, Di Tang, Wenlai Micromachines (Basel) Article Electrical impedance tomography (EIT) is a non-invasive, radiation-free imaging technique with a lot of promise in clinical monitoring. However, since EIT image reconstruction is a non-linear, pathological, and ill-posed issue, the quality of the reconstructed images needs constant improvement. To increase image reconstruction accuracy, a grey wolf optimized radial basis function neural network (GWO-RBFNN) is proposed in this paper. The grey wolf algorithm is used to optimize the weights in the radial base neural network, determine the mapping between the weights and the initial position of the grey wolf, and calculate the optimal position of the grey wolf to find the optimal solution for the weights, thus improving the image resolution of EIT imaging. COMSOL and MATLAB were used to numerically simulate the EIT system with 16 electrodes, producing 1700 simulation samples. The standard Landweber, RBFNN, and GWO-RBFNN approaches were used to train the sets separately. The obtained image correlation coefficient (ICC) of the test set after training with GWO-RBFNN is 0.9551. After adding 30, 40, and 50 dB of Gaussian white noise to the test set, the attained ICCs with GWO-RBFNN are 0.8966, 0.9197, and 0.9319, respectively. The findings reveal that the proposed GWO-RBFNN approach outperforms the existing methods when it comes to image reconstruction. MDPI 2022-07-15 /pmc/articles/PMC9322610/ /pubmed/35888936 http://dx.doi.org/10.3390/mi13071120 Text en © 2022 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
Wang, Guanghua
Feng, Di
Tang, Wenlai
Electrical Impedance Tomography Based on Grey Wolf Optimized Radial Basis Function Neural Network
title Electrical Impedance Tomography Based on Grey Wolf Optimized Radial Basis Function Neural Network
title_full Electrical Impedance Tomography Based on Grey Wolf Optimized Radial Basis Function Neural Network
title_fullStr Electrical Impedance Tomography Based on Grey Wolf Optimized Radial Basis Function Neural Network
title_full_unstemmed Electrical Impedance Tomography Based on Grey Wolf Optimized Radial Basis Function Neural Network
title_short Electrical Impedance Tomography Based on Grey Wolf Optimized Radial Basis Function Neural Network
title_sort electrical impedance tomography based on grey wolf optimized radial basis function neural network
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9322610/
https://www.ncbi.nlm.nih.gov/pubmed/35888936
http://dx.doi.org/10.3390/mi13071120
work_keys_str_mv AT wangguanghua electricalimpedancetomographybasedongreywolfoptimizedradialbasisfunctionneuralnetwork
AT fengdi electricalimpedancetomographybasedongreywolfoptimizedradialbasisfunctionneuralnetwork
AT tangwenlai electricalimpedancetomographybasedongreywolfoptimizedradialbasisfunctionneuralnetwork