Study On Image Reconstruction Algorithms For Reverse Translation Based Tomography

Computed Tomography(CT) has been one of the key technologies in medical diagnosis since the 1970 s. The existing medical CT equipment are expensive and difficult to move due to their CT system. Most of them have been applied to the major medical institutions of the developed or developing countries, however, it is hardly accessible for the small medical institutions. Therefore, it is significant and urgent to develop the image reconstruction algorithms based on the extant CT, applicable for the new CT system, which has simple structure, low cost and high mobility.In 2014, a novel CT scanning method, which is based on the reverse translation of the tube/detectors(reverse linear scanning CT for short), was proposed by Professor Ge Wang at Resselaer Polytechnic Institute of USA and Professor Fenglin Liu at Chongqing University of China. With this scheme, the popular slip ring can be replaced by a translation based setup, which is easy to set up, movable and especially low cost. Meanwhile, the novel CT scanning method has a larger scanning angular range compared with linear cone beam CT scanning, whose source/detector translates in the same direction. Motivated by the reverse linear scanning CT, this thesis put emphasis on image reconstruction algorithms with the reverse linear scanning. The simulated experiments verified the feasibility and effectiveness of the studied reconstruction algorithm, which were applied on the reverse linear scanning CT. This paper mainly includes following contents.Above all, the image reconstruction algorithm, which is Filtered Back-Projection(FBP), was studied for the reverse translation based tomography scheme with two or more linear scans. It is obvious that the projections from single reverse translation(1T) mode cannot get the data of 180?angular range. To approximate the full 180? angular range, either two orthogonal translations(2T) or three symmetric translations(3T) were used for the data acquisition scheme. The simulated experiments validated that the reconstructed result with 3T scanning modes should more resemble the original image than 2T scanning mode and be comparable with circular fan beam scanning mode using the gray histogram of the image section.Furthermore, since iterative reconstruction algorithms have the ability to reconstruct the images with high quality using less projection data, they gradually become a hot research topic and have been widely used. Therefore, iterative reconstruction algorithm based total variation minimization(TVM) was applied for the reverse translation based tomography scheme and the reverse translation based interior tomography. Simulated experiments showed that TVM based iterative reconstruction algorithm could improve the quality of reconstructed images.Finally, according to image reconstruction algorithm for the reverse translation based fan-beam tomography, slide corrected-TVM, which introduced the slide correction into TVM regularization algorithm, and TVM iterative reconstruction algorithm were utilized for reverse translation based cone beam tomography scheme. Simulated experiments were conducted to validate the feasibility of the scheme.