Feasibility of sinogram reconstruction based on inpainting method with decomposed sinusoid-like curve (S-curve) using total variation (TV) noise reduction algorithm in computed tomography (CT) imaging system: A simulation study

摘要

The use of computed tomography (CT) imaging system has increased significantly because it is very important role in the field of the medical diagnostic disease. However, CT has potential risk for high radiation dose and in addition to new strategy of reducing dose such as development of image reconstruction algorithm in sparse view conditions. Also, noise reduction is essential for improving image performance. In this study, with an aim to confirm feasibility of sinogram reconstruction based on inpainting method with decomposed sinusoid-like curve (S-curve) using total variation (TV) noise reduction algorithm in CT imaging system. For that purpose, we designed above-mentioned reconstruction and noise reduction algorithms and quantitatively evaluated coefficient of variation (COV), contrast to noise ratio (CNR) and root mean square error (RMSE). According to the results, our proposed image reconstruction method using TV noise reduction algorithm can acquire superb result in all evaluation parameters. The main benefit of our proposed method in sparse projection view is that it provides excellent image performance with efficient reconstruction in sinogram domain and noise reduction ratio. In conclusion, our results demonstrated that the better image performance using our proposed method can expect acquiring low scan time and low radiation dose. (C) 2018 Elsevier GmbH. All rights reserved.