Multiple Regression Fitting Electrical Impedance Spectro-Tomography for Quantitative Image Reconstruction of Dead Cell Fraction and Cell Concentration

Abstract

A novel image reconstruction method called multiple regression fitting electrical impedance spectro-tomography (mrf-EIST) has been proposed in order to realize the quantitative image reconstruction of dead cell fraction $\phi _{d}$ and cell concentration $c_{c}$ in a huge amount of cell environment. mrf-EIST statistically selects frequencies to extract two variables $\psi _{d}$ and $\psi _{c}$ , which quantify $\phi _{d}$ and $c_{c}$ , respectively. The $\phi _{d}$ and $c_{c}$ images are reconstructed by solving the inverse problem using $\psi _{d}$ and $\psi _{c}$ . To validate the performance of mrf-EIST, the image reconstruction by mrf-EIST in the frequency range from 100 Hz to 1 MHz is carried out under the condition that the number of cells is over 109 cells. As a result, mrf-EIST shows that the image quality defined by the difference in pixel value from the true image is less than 0.050 in $\phi _{d}$ and 0.071 in $c_{c}$ , respectively. In comparison to frequency-difference EIT (fd-EIT) as a conventional EIST regarding a position error of center of gravity, mrf-EIST provides much more accurate images, qualitatively and quantitatively, compared to the fd-EIT.

Keywords

• Electrical impedance spectro-tomography
• multiple regression fitting
• dead cell fraction
• cell concentration