Electrical impedance tomography (EIT) is a non-invasive technique for visualizing the internal structure of a human body. Capacitively coupled electrical impedance tomography (CCEIT) is a new contactless EIT technique that can potentially be used as a wearable device. Recent studies have shown that a machine learning-based approach is very promising for EIT image reconstruction. Most of the studies concern models containing up to 22 electrodes and focus on using different artificial neural network models, from simple shallow networks to complex convolutional networks. However, the use of convolutional networks in image reconstruction with a higher number of electrodes requires further investigation. In this work, two different architectures of artificial networks were used for CCEIT image reconstruction: a fully connected deep neural network and a conditional generative adversarial network (cGAN). The training dataset was generated by the numerical simulation of a thorax phantom with healthy and illness-affected lungs. Three kinds of illnesses, pneumothorax, pleural effusion, and hydropneumothorax, were modeled using the electrical properties of the tissues. The thorax phantom included the heart, aorta, spine, and lungs. The sensor with 32 area electrodes was used in the numerical model. The ECTsim custom-designed toolbox for Matlab was used to solve the forward problem and measurement simulation. Two artificial neural networks were trained with supervision for image reconstruction. Reconstruction quality was compared between those networks and one-step algebraic reconstruction methods such as linear back projection and pseudoinverse with Tikhonov regularization. This evaluation was based on pixel-to-pixel metrics such as root-mean-square error, structural similarity index, 2D correlation coefficient, and peak signal-to-noise ratio. Additionally, the diagnostic value measured by the ROC AUC metric was used to assess the image quality. The results showed that obtaining information about regional lung function (regions affected by pneumothorax or pleural effusion) is possible using image reconstruction based on supervised learning and deep neural networks in EIT. The results obtained using cGAN are strongly better than those obtained using a fully connected network, especially in the case of noisy measurement data. However, diagnostic value estimation showed that even algebraic methods allow us to obtain satisfactory results.
1 January 2021 in Computers in Biology and Medicine
Efficient electrical impedance imaging fault detection in grounding grid is necessary. In this paper, we propose an optimized differential-Multigrid-Homotopy algorithm for grounding grid impedance imaging. To enhance the iteration speed, we introduced an optimized differential algorithm to obtain the Jacobian matrix and derive calculation methods for the two-dimensional constraint operator and interpolation operator. Moreover, we utilize multigrid fast iteration and homotopy methods simultaneously to improve the rapidity and ill-posedness of the inverse problem. We validate the proposed algorithm through numerical simulations and experimental studies. First, we verify the time and convergence of the optimized differential algorithm for solving the Jacobian matrix. Secondly, we investigate the CPU time and convergence of image reconstruction based on the Tikhonov algorithm, HT algorithm, and Multigrid-Homotopy (MG-HT) algorithm under different ground grid fault scenarios and noise levels. The simulation and experimental results show that the optimized differential-Multigrid-Homotopy algorithm has better robustness and convergence.
Magnetically coupled resonant technology is a novel method for solving the breakpoint locating of power grounding grid. But the method can only detect breakpoints of a single mesh grounding grid at present. In this paper, a magnetically coupled resonant detection method for four-hole grounding grid breakpoint is proposed. Firstly, the equivalent circuit model of the four mesh grounding grid with two types of breakpoints, namely edge branch and intermediate branch, is established. The input impedance and phase angle of the system are obtained by analyzing the equivalent capacitance and equivalent resistance in the model. Secondly, the magnetically coupled resonant physical process of grounding grid faults is solved via HFSS software. The magnetic field intensity and phase frequency characteristic curves of four mesh holes with different branches and positions of breakpoints and different corrosion degrees are studied, and an experimental system is built to verify the feasibility. The results show that under the condition of grounding grid buried depth of 0.5 m and input frequency of 1∼15MHz, and there is an inverse relationship between equivalent capacitance and distortion frequency, the phase angle is positively correlated with the degree of corrosion of grounding grid, and the error of signal distortion frequency can be positioned at 5%. This paper provides some ideas for the application of magnetic coupling grounding grid detection technology.