Loading...

Medical Microwave Imaging using a Priori Information

Hosseini, Mohammad | 2025

0 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 58269 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Shishegar, Amir Ahmad
  7. Abstract:
  8. This dissertation introduces a method for reducing errors in prior information related to the actual geometry of biological tissues in microwave imaging. These errors, often caused by boundary mismatches or inaccurate modeling of tissue heterogeneity, significantly impact the accuracy of retrieved tissue dielectric properties. To address this issue, we propose a shape optimization-based approach that aligns prior boundary data with actual tissue boundaries using measured microwave imaging data. Our method features an efficient technique for computing the sensitivity of scattered electric fields to node positions. By deforming the mesh structure without altering its topology, this approach improves the alignment between prior models and the imaged tissue. Simulations show that even under severe noise conditions, the Hausdorff distance error between prior information and true tissue boundaries can be reduced by up to 80%. Additionally, unlike previous studies that focus solely on spatial and boundary information, this work incorporates additional prior data—such as MRI intensity values, typically overlooked in conventional methods—for dielectric property reconstruction. We propose a DBIM-based approach to establish a mapping between MRI intensity data and tissue dielectric properties, leading to a more accurate representation of tissue heterogeneity. Simulations demonstrate that integrating boundary correction with this technique enhances both boundary alignment and dielectric property recovery
  9. Keywords:
  10. Microwave Imaging ; Shape Optimization ; Image Registration ; Distorted Born Iterative Method (DBIM) ; Registration Algorithm ; A Priori Information ; Medical Imaging

 Digital Object List

 Bookmark

...see more