In contrast to the conventional wisdom that the interior problem does not have a unique solution, in 2007 we developed and patented the interior tomography technology (exact and stable interior reconstruction from x-ray projection data associated with only lines through an internal region of interest). When applying conventional CT algorithms to such an incomplete dataset, the features outside the ROI may create artifacts overlapping inside real features. Our new interior reconstruction theory assumes exact knowledge on a small subregion in the region of interest, such as air in the lungs, blood in the heart, etc. As a surprising result, exact and stable interior reconstruction becomes feasible! This break though has numerous biomedical implications, and will benefit or enable a wide range of CT applications including lung CT, cardiac CT, temporal bone imaging, intraoperative CT and so on, where we need to handle large objects, minimize radiation dose, suppress scattering artifacts, enhance temporal resolution, reduce system cost, and increase scanner throughput. The overall goal of this project is to develop both analytic and iterative reconstruction algorithms and software for interior tomography.
The specific aims are to (1) develop analytic and theoretically exact interior reconstruction algorithms that have not been possible before;(2) develop efficient iterative interior reconstruction algorithms that incorporate additional constraints to improve image quality and reduce radiation dose further;and (3) evaluate the proposed interior reconstruction algorithms in numerical simulation and phantom experiments and develop the first of its kind software package for interior tomography. On completion of this project, we will have developed both analytic and iterative exact interior reconstruction algorithms for preclinical and clinical CT applications, and develop a commercial software package ready to be competitive on the marketplace.
In 2007 we developed and patented the interior tomography technology (exact and stable interior reconstruction from x-ray projection data associated with only lines through an internal region of interest). Our theory assumes exact knowledge on a small subregion in the region of interest, such as air in the lungs, blood in the heart, etc., so that for the first time exact and stable interior reconstruction becomes realistic. Our proposed methods and software will benefit or enable a wide range of CT applications including lung CT, cardiac CT, temporal bone imaging, intraoperative CT and so on, where we need to handle large objects, minimize radiation dose, suppress scattering artifacts, enhance temporal resolution, reduce system cost, and increase scanner throughput.
| Ye, Yangbo; Yu, Hengyong; Wang, Ge (2011) Gel'fand-Graev's reconstruction formula in the 3D real space. Med Phys 38 Suppl 1:S69 |
| Yu, Hengyong; Wang, Ge (2010) A soft-threshold filtering approach for reconstruction from a limited number of projections. Phys Med Biol 55:3905-16 |
| Wang, Ge; Yu, Hengyong (2010) Can interior tomography outperform lambda tomography? Proc Natl Acad Sci U S A 107:E92-3, author reply E94-5 |
