Optimizing Image Quality and Diagnostic Accuracy in Impedance Tomography with Gastric Phantom Model

The phantom model imaging system enables real-time data processing and visualization through Electrical Impedance Tomography (EIT). In gastric impedance tomography, a 16 of electrodes placed around the phantom measure electrical impedance, thereby monitoring the distribution and motility of gastric contents. Image reconstruction aims to obtain an optimal range of image values to ensure superior image quality. The process involves reconstructing the mesh area for analysis and determining computational speed in image reconstruction, with a mesh value of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$5,5\times 10^{-5}$</tex> pixels chosen as the basis for determining the resolution level in the hyperparameters lambVal <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(\Lambda)$</tex> and p Val <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(p)$</tex>. The hyperparameter <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$p$</tex> is utilized to find the optimal value for the EIT image shape that approximates the boundary ground truth with the smallest pixel error, while <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\Lambda$</tex> is optimized for detail and accuracy of the image scale factor. The diagnostic system's performance was evaluated in terms of accuracy, response time, and measurement reliability, using a stable testing frequency of 50,000. A frequency of 50,000 was selected due to the decrease and stabilization of the relative permittivity value, leading to an error of 0.36555. Validation of image reconstruction results is conducted using both qualitative and quantitative methods. The qualitative approach involves comparing the real image boundary with its convergence range, producing a convergence area within the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\rho$</tex> range of 0.36 to 0.48 and the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\Lambda$</tex> range of 0.63 to 0.84. The quantitative approach employs statistical techniques and numerical measurements for more objective validation which yields a convergence area with a <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$p$</tex> of 0.45 and a <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\Lambda$</tex> of 0.7 at a frequency of 50,000 Hz. Additionally, real-time processing enhances the system's capability to monitor and adapt to rapid changes in the gastric environment, further improving diagnostic accuracy and the effectiveness of the imaging system.