Discrete Dipole Approximation-Based Microwave Tomography for Fast Breast Cancer Imaging

Area of Science:

• Electromagnetics and Applied Physics
• Biomedical Engineering
• Computational Imaging

Background:

• Microwave tomography is a promising imaging modality, but reconstruction speed has been a limitation.
• Existing algorithms often require substantial computational resources and time.
• The discrete dipole approximation (DDA) offers a potential solution for efficient forward modeling in microwave tomography.

Purpose of the Study:

• To develop and validate a fast microwave tomography reconstruction algorithm.
• To leverage the two-dimensional discrete dipole approximation (2D-DDA) as an efficient forward solver.
• To demonstrate the algorithm's capability in reconstructing images from both synthetic and experimental data.

Main Methods:

• The algorithm employs the 2D-DDA for forward modeling, suitable for low-profile antennas and lossy media.
• A microwave measurement system with 16 monopole antennas and a vector network analyzer was utilized.
• The nodal adjoint method was used for efficient Jacobian matrix computation, enabling rapid image reconstruction.

Main Results:

• The algorithm successfully reconstructed a 2D plane of a cylindrical phantom using both synthetic and experimental data.
• Image reconstruction was achieved in under 6 seconds, demonstrating significant time savings.
• The 2D-DDA approach proved effective given the low-profile antennas and lossy coupling liquid.

Conclusions:

• The developed microwave tomography algorithm offers a substantial improvement in reconstruction speed and efficiency.
• The combination of 2D-DDA and nodal adjoint method is highly effective for fast microwave imaging.
• This approach presents a significant advancement over previous microwave tomography implementations, reducing time and memory requirements.