Experimental evaluation of two iterative reconstruction methods for induced current electrical impedance tomography

Area of Science:

• Electrical Impedance Tomography
• Biomedical Imaging

Background:

• Induced Current Electrical Impedance Tomography (ic-EIT) offers a novel approach for imaging conductivity distributions.
• Reconstruction algorithms are crucial for accurate image generation in ic-EIT.

Purpose of the Study:

• To compare the performance of two regularization methods, Levenberg-Marquardt (N-R-LM) and Hachtel's Augmented Matrix (N-R-HAM), for image reconstruction in ic-EIT.
• To evaluate the resolution and noise immunity of the N-R-HAM method for conductivity imaging.

Main Methods:

• Utilized the Newton-Raphson (N-R) iterative method with N-R-LM and N-R-HAM regularization.
• Employed experimental data from a cylindrical tank with 16 electrodes and 8 surrounding coils driven by a 20-kHz sine wave.
• Computed scalar potential fields using the Finite Element Method.

Main Results:

• The N-R-HAM method demonstrated superior resolution and noise immunity compared to the N-R-LM method.
• Successfully resolved two non-conducting rods within the cylindrical tank.
• Achieved a spatial resolution of approximately 15% of the tank diameter.

Conclusions:

• The N-R-HAM method is a promising technique for high-resolution conductivity imaging in ic-EIT.
• Increasing the number of coils and/or electrodes could further enhance spatial resolution.
• Stray magnetic flux presents a challenge for static conductivity imaging in ic-EIT.