Real-Time 3D Microwave Medical Imaging With Enhanced Variational Born Iterative Method.

Area of Science:

• Electromagnetics
• Medical Imaging
• Computational Physics

Background:

• Microwave imaging (MWI) systems require efficient algorithms for real-time reconstruction of material properties.
• Existing methods like the Born iterative method (BIM) and distorted Born iterative method (DBIM) can be computationally intensive.
• Accurate reconstruction of dielectric permittivity and conductivity is crucial for MWI applications.

Purpose of the Study:

• To introduce a novel variational Born iterative method (VBIM) for accelerated real-time microwave imaging.
• To integrate the real and imaginary separation (RIS) approach for simultaneous dielectric property reconstruction.
• To demonstrate the computational efficiency and accuracy of the VBIM-RIS method.

Main Methods:

• Implementation of S-parameter volume integral equation and waveport vector Green's function.
• Application of the real and imaginary separation (RIS) approach at each iterative step.
• Utilization of graphics processing unit (GPU) acceleration for real-time performance.

Main Results:

• VBIM demonstrates reduced computational time compared to BIM and DBIM.
• Successful reconstruction of dielectric permittivity and conductivity using the VBIM-RIS method.
• Real-time imaging achieved with approximately 5 seconds per reconstruction frame for a brain-tumor phantom.

Conclusions:

• The VBIM-RIS method offers a computationally efficient and accurate solution for real-time microwave imaging.
• The algorithm is validated through synthetic data and experimental measurements.
• This method shows significant potential for monitoring procedures like thermal therapy in real-time.