Direct EIT Jacobian calculations for conductivity change and electrode movement

Area of Science:

• Medical Imaging
• Computational Electromagnetics
• Biomedical Engineering

Background:

• Electrical impedance tomography (EIT) is sensitive to boundary shape deformations, causing artifacts in lung imaging due to patient movement.
• Current EIT reconstruction methods often struggle with accuracy and speed, especially for dynamic 2D and 3D models.
• Calculating the 'movement Jacobian' is crucial but computationally intensive with existing perturbation techniques.

Purpose of the Study:

• To develop a faster and more accurate method for calculating the movement Jacobian in EIT.
• To overcome the limitations of perturbation methods in handling large finite element models for deformable media.
• To improve the feasibility of EIT for dynamic lung imaging applications.

Main Methods:

• A direct method for calculating the Jacobian was formulated using derivatives of the finite element system matrix with respect to geometry changes.
• The proposed direct method was compared against traditional perturbation techniques.
• The computational efficiency and accuracy of both methods were evaluated using illustrative examples.

Main Results:

• The proposed direct method demonstrated a significant speed improvement, being approximately 300 times faster than perturbation methods.
• The direct method maintained accuracy for larger model sizes, whereas perturbation methods began to diverge.
• This advancement addresses limitations that previously restricted deformable EIT reconstructions primarily to 2D.

Conclusions:

• The direct Jacobian calculation method offers a substantial advancement for EIT, particularly for dynamic and deformable imaging scenarios like lung monitoring.
• This technique enhances computational efficiency and accuracy, paving the way for more robust 2D and potentially 3D EIT applications.
• The improved method could lead to more reliable and widespread use of EIT in clinical settings.