A New Computational Method for Large-Scale Inverse Problems in Magnetic Induction Tomography
Mentor 1
Amhed Kaffel
Start Date
28-4-2023 12:00 AM
Description
Magnetic Induction Tomography (MIT) is a medical imaging modality that offers low-cost and portability advantages over other imaging modalities such as MRI and CT scans. Many successful regularization methods employed to solve large scale linear inverse problems in imaging applications (such as image deblurring, image inpainting, and computed tomography) aim to enhance edges in the solution. The Singular Value Decomposition (SVD)-based image reconstruction algorithm in MIT faces an ill-posed problem and has shown some limitations. In this study, we will develop a new, efficient image reconstruction algorithm to improve the accuracy and reliability of MIT. We aim to address the issues of the SVD-based image reconstruction algorithm in MIT by proposing a new hybrid method based on the combination of the Tikhonov regularization technique with an iterative method, which is efficient in solving ill-posed integral equations. The hybrid method is expected to enhance computational speed and improve reconstruction accuracy for large-scale ill-posed problems. The scanning process in MIT will be explained, and the shortcomings of the current algorithm will be analyzed. To evaluate the performance of the proposed method, the study will investigate its results through numerical examples. Comparative analysis of the method's performance will be conducted, and the results will show significant improvements in terms of accuracy and speed of the new hybrid method over the existing SVD-based method. Therefore, the proposed hybrid method has the potential to make MIT a primary imaging modality in medical applications, as it can provide more accurate and reliable results. We will also introduce and incorporate other methods, like multiwavelets and deep learning, that the team will use to overcome these barriers. Since this is still an ongoing research project, the team will continue to work on improving the results. The end goal is to apply this study to real medical cases.
A New Computational Method for Large-Scale Inverse Problems in Magnetic Induction Tomography
Magnetic Induction Tomography (MIT) is a medical imaging modality that offers low-cost and portability advantages over other imaging modalities such as MRI and CT scans. Many successful regularization methods employed to solve large scale linear inverse problems in imaging applications (such as image deblurring, image inpainting, and computed tomography) aim to enhance edges in the solution. The Singular Value Decomposition (SVD)-based image reconstruction algorithm in MIT faces an ill-posed problem and has shown some limitations. In this study, we will develop a new, efficient image reconstruction algorithm to improve the accuracy and reliability of MIT. We aim to address the issues of the SVD-based image reconstruction algorithm in MIT by proposing a new hybrid method based on the combination of the Tikhonov regularization technique with an iterative method, which is efficient in solving ill-posed integral equations. The hybrid method is expected to enhance computational speed and improve reconstruction accuracy for large-scale ill-posed problems. The scanning process in MIT will be explained, and the shortcomings of the current algorithm will be analyzed. To evaluate the performance of the proposed method, the study will investigate its results through numerical examples. Comparative analysis of the method's performance will be conducted, and the results will show significant improvements in terms of accuracy and speed of the new hybrid method over the existing SVD-based method. Therefore, the proposed hybrid method has the potential to make MIT a primary imaging modality in medical applications, as it can provide more accurate and reliable results. We will also introduce and incorporate other methods, like multiwavelets and deep learning, that the team will use to overcome these barriers. Since this is still an ongoing research project, the team will continue to work on improving the results. The end goal is to apply this study to real medical cases.
