An image reconstruction framework based on deep neural network for electrical impedance tomography

Li, Xiuyan; Lu, Yang; Wang, Jianming; Dang, Xin; Wang, Qi; Duan, Xiaojie; Sun, Yukuan

doi:10.1109/icip.2017.8296950

Cited by 26 publications

(12 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It can efficiently learn high-level features from data through a hierarchical network framework [12]. Several deep learning algorithms have been proposed for image reconstruction of electrical tomography using different network models [13]- [16]. Jin et al established a benchmark dataset for Electrical Capacitance Tomography and put forward a deep autoencoder-based iteration method to evaluate reconstruction results on this database [13].…”

Section: Gas-mentioning

confidence: 99%

“…The DELM combined with split Bregman (SB) method and fast iterative shrinkage thresholding (FIST) algorithm realized competitive reconstruction both on numerical and experimental data. Li et al presented a deep neural network combined with stacked autoencoder and logistic regression to improve the quality of reconstruction in Electrical Impedance Tomography [16]. However, most of such studies focus on the image reconstruction of different inclusions, and in such cases all the electrodes are in contact with conductive fluid.…”

Section: Gas-mentioning

confidence: 99%

See 1 more Smart Citation

Gas-Liquid Two-Phase Stratified Flow Interface Reconstruction With Sparse Batch Normalization Convolutional Neural Network

Tan

et al. 2021

IEEE Sensors J.

View full text Add to dashboard Cite

Two-phase stratified flow is ubiquitous in industrial processes, monitoring its phase interface is important for improving the safety and efficiency of process. Electrical Resistance Tomography is a promising non-intrusive visualization technique for monitoring two-phase flow. However, some electrodes could lose their contact with the liquid in stratified flow, aggravating the under-determined image reconstruction solution and resulting in low-quality reconstructed image in traditional methods. In this paper, a sparse batch normalization convolutional neural network (SBN-CNN) method is proposed for accurate and rapid gas-liquid interface reconstruction. Up-sampling and normalization were used for preprocessing of measurements. A novel network structure including convolution, pooling and batch normalization layers was designed to extract features from sparse measurements and the images were reconstructed by deep fully connected layers. After batch-based training, the proposed SBN-CNN achieves a superior performance to the-state-of-art methods in terms of convergence rate, imaging accuracy, noise resistance and generalization ability, through experimental validation.

show abstract

Section: Gas-mentioning

confidence: 99%

Section: Gas-mentioning

confidence: 99%

Gas-Liquid Two-Phase Stratified Flow Interface Reconstruction With Sparse Batch Normalization Convolutional Neural Network

Tan

et al. 2021

IEEE Sensors J.

View full text Add to dashboard Cite

show abstract

“…Among the regularization approaches for the nonlinear inverse problem, the "D-bar method" [301] and the "Calderón method" [302] are proven strategies which have been extended in [303] with a convolutional neural network to prevent image blurring. Deep learning approaches applied to image reconstruction for EIT have also been addressed in [304,305]. The incorporation of structural information has been applied in [306] to reduce the dimensionality and sensitivity to regularization parameters.…”

Section: Electrical Impedance Tomographymentioning

confidence: 99%

Fundamentals, Recent Advances, and Future Challenges in Bioimpedance Devices for Healthcare Applications

2019

View full text Add to dashboard Cite

This work develops a thorough review of bioimpedance systems for healthcare applications. The basis and fundamentals of bioimpedance measurements are described covering issues ranging from the hardware diagrams to the configurations and designs of the electrodes and from the mathematical models that describe the frequency behavior of the bioimpedance to the sources of noise and artifacts. Bioimpedance applications such as body composition assessment, impedance cardiography (ICG), transthoracic impedance pneumography, electrical impedance tomography (EIT), and skin conductance are described and analyzed. A breakdown of recent advances and future challenges of bioimpedance is also performed, addressing topics such as transducers for biosensors and Lab-on-Chip technology, measurements in implantable systems, characterization of new parameters and substances, and novel bioimpedance applications.

show abstract

“…Their method of evaluating resolution is also used for reference by other scholars. The stacked auto encoder (SAE) is an old method similar to one-dimensional convolution before CNN, also proved to be an effective approach to enhance images lately [ 29 ]. In 2019, Chen et al [ 30 ] proposed an algorithm based on an SAE neural network to reconstruct the anomaly in biological tissues.…”

Section: Introductionmentioning

confidence: 99%

Application of a Generative Adversarial Network in Image Reconstruction of Magnetic Induction Tomography

Yang

Liu

Wang

et al. 2021

Sensors

View full text Add to dashboard Cite

Image reconstruction of Magnetic induction tomography (MIT) is an ill-posed problem. The non-linear characteristics lead many difficulties to its solution. In this paper, a method based on a Generative Adversarial Network (GAN) is presented to tackle these barriers. Firstly, the principle of MIT is analyzed. Then the process for finding the global optimum of conductivity distribution is described as a training process, and the GAN model is proposed. Finally, the image was reconstructed by a part of the model (the generator). All datasets are obtained from an eight-channel MIT model by COMSOL Multiphysics software. The voltage measurement samples are used as input to the trained network, and its output is an estimate for image reconstruction of the internal conductivity distribution. The results based on the proposed model and the traditional algorithms were compared, which have shown that average root mean squared error of reconstruction results obtained by the proposed method is 0.090, and the average correlation coefficient with original images is 0.940, better than corresponding indicators of BPNN and Tikhonov regularization algorithms. Accordingly, the GAN algorithm was able to fit the non-linear relationship between input and output, and visual images also show that it solved the usual problems of artifact in traditional algorithm and hot pixels in L2 regularization, which is of great significance for other ill-posed or non-linear problems.

show abstract

An image reconstruction framework based on deep neural network for electrical impedance tomography

Cited by 26 publications

References 10 publications

Gas-Liquid Two-Phase Stratified Flow Interface Reconstruction With Sparse Batch Normalization Convolutional Neural Network

Gas-Liquid Two-Phase Stratified Flow Interface Reconstruction With Sparse Batch Normalization Convolutional Neural Network

Fundamentals, Recent Advances, and Future Challenges in Bioimpedance Devices for Healthcare Applications

Application of a Generative Adversarial Network in Image Reconstruction of Magnetic Induction Tomography

Contact Info

Product

Resources

About