An improved Wexler algorithm for electrical impedance tomography using finite element method and gradient based...

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Computational Science

Background:

• Electrical impedance tomography (EIT) requires efficient reconstruction algorithms for real-time applications.
• The increasing popularity of wearable technology necessitates advancements in EIT.

Purpose of the Study:

• To enhance the historical Wexler scheme for EIT reconstruction.
• To improve the speed, reliability, and image quality of EIT.

Main Methods:

• Adapted the Wexler scheme for finite element method (FEM) implementation.
• Introduced overrelaxation and a variable step size dependent on the objective function's gradient for improved convergence.
• Developed a simplified electrode model to enhance image quality.
• Proposed a method for determining initial conductivity.

Main Results:

• The improved Wexler scheme achieves performance comparable to standard algorithms like NOSER and Gauss-Newton with total variational regularization.
• Demonstrated the algorithm's scalability for large-mesh EIT tests.
• Enhanced image quality through a new electrode model.

Conclusions:

• The modified Wexler scheme offers a competitive and scalable solution for EIT reconstruction.
• These improvements are crucial for advancing wearable EIT devices and applications.