A Deep Learning-Based EIT System for Robust Gesture Recognition Under Confounding Factors.

Electrical Impedance Tomography (EIT) is a noninvasive imaging technique that utilizes electrical voltage data to reconstruct cross-sectional images of the human body. EIT has diverse applications such as brain imaging and gesture recognition, making it a valuable tool for both medical diagnostics and human-computer interaction. One potential drawback of EIT is the power consumption and computational requirements for image reconstruction, which may limit real-time applications. This paper presents a novel approach that directly applies machine learning to raw EIT voltage data, bypassing the image reconstruction phase to achieve faster and highly accurate gesture classification. We introduce the Statistical Analysis, Information Theory, and Data-Driven (SID) pipeline, which analyzes raw EIT data using statistical techniques, information theory, and feature ranking, followed by classification with machine learning models. Two datasets were collected for this study: one for five gesture recognition (1193 samples) and another for distinguishing between arm flexion and extension (719 samples). The proposed SID pipeline was applied where the gesture recognition and arm flexion/extension feature set was reduced from 40 to 2 and 1, respectively, and XGBoost was used for classification and achieved an accuracy of 90.38% and 100.00%, respectively. The proposed method demonstrated high accuracy in both tasks while using a reduced feature set. This is in addition to by-passing the image reconstruction for EIT, further enhancing computational efficiency and reduced power consumption, highlighting the potential of this approach for real-time applications.

Electrical brain imaging (EBI) refers to a set of techniques that exploit either the spontaneous electrical activity of the central nervous system, as in electroencephalographic (EEG) source reconstruction, or make use of external current injections, as in electrical impedance tomography (EIT) , to image the structure or function of the brain. When compared to other brain imaging methods used in research or in the clinical setting, such as computed tomography (CT), magnetic resonance imaging (MRI), functional MRI (fMRI), positron emission tomography (PET) and single photon emission computed tomography (SPECT), EIT and EEG source localization instrumentation offer the advantages of portability, low cost, high temporal resolution [ms] and quick setup. The downsides are a low spatial resolution [cm], high computational cost of the image reconstruction process and high sensitivity to imperfections of the electrical model of the head.
\nIn this work, a new special purpose reconstruction algorithm for EIT is presented and validated wth experimental measurements performed on a cylindrical phantom and on a simulated human head. The algorithm focuses on the quick detection of compact conductivity contrasts in imperfectly known in 3D domains. The performance of the proposed algorithm is then compared to the one of a benchmark reconstruction method in the EIT field, Tikhonov regularized reconstruction, with stroke detection and classification as a case study. Moreover, the possible application of EIT imaging to the detection of epileptic foci with intracranial deep electrodes (stereoelectroencephalography or SEEG) is explored.
\nFinally, EEG source reconstruction algorithms are implemented on a heterogeneous multi-CPU and multi-GPU computing system to significantly reduce the reconstruction time.

Electrical impedance tomography (EIT) is a medical imaging technique based on the injection of a current or voltage pattern through electrodes on the skin of the patient, and on the reconstruction of the internal conductivity distribution from the voltages collected by the electrodes. Compared to other imaging techniques, EIT shows significant advantages: it does not use ionizing radiation, is non-invasive and is characterized by high temporal resolution. Moreover, its low cost and high portability make it suitable for real-time, bedside monitoring. However, EIT is also characterized by some technical limitations that cause poor spatial resolution. The possibility to design wearable devices based on EIT has recently given a boost to this technology. In this paper we reviewed EIT physical principles, hardware design and major clinical applications, from the classical to a wearable setup. A wireless and wearable EIT system seems a promising frontier of this technology, as it can both facilitate making clinical measurements and open novel scenarios to EIT systems, such as home monitoring.

This paper presents a high accuracy hand gesture recognition system based on electrical impedance tomography (EIT). The system interfaces the forearm using a wrist wrap with embedded electrodes. It measures the inner conductivity distributions caused by bone and muscle movement of the forearm in real-time and passes the data to a deep learning neural network for gesture recognition. The system has an EIT bandwidth of 500 kHz and a measured sensitivity in excess of 6.4 Ω per frame. Nineteen hand gestures are designed for recognition, and with the proposed round robin sub-grouping method, an accuracy of over 98% is achieved.

Electrical impedance tomography (EIT) imaging modality has great potential on industrial applications with the advantages of being high temporal resolution. It is especially useful in cases, such as geophysical detection, landmine detection, and detections on non-transparent region, where the measurement data are only available from single surface, for data acquisition. Instead of the circular EIT model that uses the traditional circular electrode model, in this paper, planar array EIT is implemented, aiming to visualize a pipeline transporting a two-phase flow. The planar array can explore spatial information within its detectable region by producing 3D images, which have a higher spatial resolution in the axis direction than a traditional EIT with a dual-plane electrode sensor. However, in solving the inverse problem of a 3D subsurface EIT using a planar array, the images may be degraded, especially in cases where the location of the target is relatively deep. The total variation (TV) algorithm as block prior assumption-based regularization method has the potential to improve the image quality, and some works have shown that TV reconstructs sharper images, which provides an advantage when representing spatial information. In this paper, the performance of subsurface EIT using the TV algorithm for 3D visualization is presented based on simulations and experiments, and the results of quantitative measurement of depth are discussed.

Electrical impedance tomography (EIT) is an imaging technique that reconstructs the conductivity distribution in the interior of an object using electrical measurements from the electrodes that are attached around the boundary. The Complete Electrode Model (CEM) accurately incorporates the electrode size, shape, and effective contact impedance into the forward problem for EIT. In this work, the effect of the conductivity distribution and the electrode contact impedance on the solution of the forward problem is addressed. In particular, the sensitivity of the electric potential with respect to a small-amplitude perturbation in the conductivity, and with respect to some defective electrodes is studied. The Gâteaux derivative is introduced as a tool for the sensitivity analysis and the Gâteaux differentiability of the electric potential with respect to the conductivity and to the contact impedance of the electrodes is proved. The derivative is then expressed as the unique solution to a variational problem and the discretization is performed with Finite Elements of type P1. Numerical simulations for different 2D and 3D configurations are presented. This study illustrates the impact of the presence of perturbations in the parameters of CEM on EIT measurements. Finally, the 2D inverse conductivity problem for EIT is numerically solved for some configurations and the results confirm the conclusions of the numerical sensitivity analysis.

Electrical Impedance Tomography (EIT) is an imaging technique which relies on the accuracy of boundary voltage measurement. In application to the head of EIT, the scalp voltage is very small and sensitive to noise and error. Electrode placement error is an error which is difficult to avoid in practice. Furthermore, the exact value of electrode contact impedance is difficult to know. In this study, the influence of electrode placement error and inaccurate electrode contact impedance were investigated by simulation. The result shows that the scalp voltage error due to the placement error was between 0.003 mV and 4.670 mV. Comparing to the scalp voltage response obtained from brain stimulation, which has been reported less than $90 \mu \text{V}$, the most of the error values were much larger. The voltage errors due to inaccurate contact impedance were less than $1.2 \mu \text{V}$ which is very small and ignorable. This indicates the importance of the attachment procedure of electrodes that needs to be very careful.

A time-dependent, flexible, 2-dimensional fine thoracic Finite-Element model for continuous lung Electrical Impedance Tomography (EIT) imaging is presented. It can be used as a tool to identify lung EIT hardware specifications, such as minimum frame rate, number of electrodes and noise requirements, as well as for prior evaluation of EIT reconstruction algorithms, before applying in-vivo data. The model is parameterized with respect to electrodes’ characteristics, the measuring protocol, structural and temporal properties of the thoracic tissues. It is implemented in FEMM using a script generated by MATLAB code, while the measurements are acquired also by MATLAB. The electrodes’ contact impedances, as well as the changes in electrical and geometric properties of the thoracic tissues due to breathing and perfusion, during each particular measuring frame, are taken into account.

A gesture recognition system using stereo vision and arm model fitting

Electrical impedance tomography (EIT) is an imaging technique with advantages of non-intrusiveness, low-cost and high temporal resolution, which is promising for multiphase flow instrumentation. However, it produces smooth images with low spatial resolution where the interface between phases cannot be distinguished and from which the phase fraction cannot be estimated correctly. In this article, an eigenvalue analysis of EIT raw data is used to estimate the void fraction, i.e. the phase area ratio in 2D, without reconstructing images. For a given EIT sensor, each acquisition frame is represented by an impedance matrix whose eigenvalues are computed after normalization. The main characteristics of the eigenvalue distribution for different two-phase flow patterns within a cylindrical pipe are analyzed numerically. The behaviors of the leading eigenvalue and of the sum of the absolute values of the following ones are assessed as functions of the void fraction. This leads to an estimation of the two-phase flow void fraction based on the EIT sensor configuration. The presented numerical results highlight the existing correlation between the eigenvalues and the void fraction for the phase distribution patterns considered. These simulation results are compared with experimental static tests for validation.

Electrical impedance tomography (EIT) is a non-radiative, inexpensive technique that can facilitate real-time dynamic monitoring. It has the potential to be of considerable clinical value playing an important role in diagnostics and monitoring of a number of disease conditions. EIT has been under development for over 100 years for industrial applications. Approximately 40 years ago it started to be applied to clinical problems. Some researchers tried to develop EIT as a replacement for image modality liker MRI or CT; however the really potential of EIT is its ability to monitor physiological condition related to impedance change. This is due to EITs high temporal resolution. This chapter introduces some of the basic concepts of EIT along with a number of example clinician applications, specifically, hardware, reconstruction algorithms and examples of its potential use in the brain, breast and lung.

The inverse problem solution in electrical impedance tomography (EIT) is sensitive to measurement noise. Due to modeling errors and contact impedance, EIT measurement data are often contaminated by noise, which affects the accuracy of image reconstruction. This study introduces a novel denoising method combining Discrete Wavelet Transform (DWT) with a Denoising Autoencoder (DAE) to effectively address measurement noise. The DWT-DAE method was systematically evaluated on various EIT electrode pairs data, including simulated data with different SNR levels (20-60 dB), water tank measurement data, and KIT4 public dataset. Results show that the DWTDAE method significantly reduces noise in EIT measurement data, improving the SNR from approximately 20-30 dB to over 55 dB, thereby enhancing the quality of reconstructed images. To validate the practical efficacy of our approach, we applied it to the challenging case of intact maize ears. The DWT-DAE method was used to denoise the EIT measurements, followed by image reconstruction using the Convex-Concave Procedure-based Block Sparse Bayesian Learning Algorithm (CCPBSBL). Results indicate that this integrated methodology substantially enhances the accuracy of reconstructed conductivity distributions in maize ears, demonstrating its practical value for agricultural applications.

In this paper, a multi-circle planar electrical impedance tomography (EIT) sensor is proposed for three-dimensional (3D) miniature imaging. There are two problems in 3D miniature EIT imaging: 1) the spatial resolution is low due to the number of electrodes is limited; 2) contact impedance influences image reconstruction quality severely. The proposed EIT sensor made by a printed circuit board (PCB) adopts multi-circle structure and tackles the two problems. Driven electrodes are distributed on the outermost two circles, while measurement electrodes are distributed on the innermost circle. Contact impedance is reduced due to the area of driven electrodes could be larger than traditional ones by using the multi-circle structure, it’s also owning to gold is deposited on the surface of the electrodes. Moreover, a new driven-measurement pattern of the EIT sensor is proposed to improve the spatial resolution as well as image reconstruction quality. The proposed EIT sensor is verified by both simulation and experiment. Average Image correlation coefficient (ICC) is as high as 0.8475 in simulation. Compared with the 16-electrode EIT sensor, ICC is improved 39.16%. In experiment, average ICC of four standard objects is 0.8389. The proposed 3D EIT sensor is good at miniature imaging.

A review of electrical impedance tomography in lung applications: Theory and algorithms for absolute images

This paper presents a pure element-free Galerkin method (EFGM) forward model for image reconstruction in 2D and 3D electrical impedance tomography (EIT) using an adaptive current injection method. In EIT systems with the adapting current injection method, both static and dynamic images can be reconstructed; however, determination of electrode contact impedances in the complete electrode model is difficult and the Gap model is used. In this paper, in the EIT forward problem a weak form functional based on the Gap model and a pure EFGM approach are developed, and in the EIT inverse problem, Jacobian matrix is computed by the EFGM, and a fast integration technique is introduced to calculate the entries of the Jacobian matrix within an adequate computation time. The influence of increasing the density of nodes at and near the electrodes with steep electric potential gradients on the accuracy of FEM and EFGM forward solutions is investigated, and the performance of the image reconstruction algorithm with the proposed fast integration technique is examined. The numerical results reveal that the proposed EFGM forward model with the fast integration technique has an efficient performance both in terms of mean relative imaging errors and computational time.