版权所有:内蒙古大学图书馆 技术提供:维普资讯• 智图
内蒙古自治区呼和浩特市赛罕区大学西街235号 邮编: 010021
作者机构:UCL Inst Nucl Med London NW1 2BU England Yale Univ Yale PET Ctr New Haven CT 06520 USA Univ Bretagne Occidentale LaTIM INSERM UMR 1101 F-29238 Brest France UCL Dept Comp Sci London WC1E 6BT England GE Healthcare MICT Engn Waukesha WI 53188 USA
出 版 物:《IEEE TRANSACTIONS ON RADIATION AND PLASMA MEDICAL SCIENCES》 (IEEE Trans. Radiat. Plasma Med. Sci.)
年 卷 期:2021年第5卷第3期
页 面:362-372页
核心收录:
基 金:GE Healthcare National Institute for Health Research, University College London Hospitals Biomedical Research Centre EPSRC [EP/H046410/1, EP/T026693/1, EP/M022587/1, EP/K005278/1, EP/P022200/1] Funding Source: UKRI
主 题:Attenuation Image reconstruction Splines (mathematics) Strain Linear programming Estimation Computed tomography Anatomical prior misalignment estimation penalized image reconstruction
摘 要:Two algorithms for solving misalignment issues in penalized PET/CT reconstruction using anatomical priors are proposed. Both approaches are based on a recently published joint motion estimation and image reconstruction method. The first approach deforms the anatomical image to align it with the functional one while the second approach deforms both images to align them with the measured data. Our current implementation alternates between alignment estimation and image reconstruction. We have chosen parallel level sets (PLSs) as a representative anatomical penalty, incorporating a spatially variant penalty strength. The performance was evaluated using simulated nontime-of-flight data generated with an XCAT phantom in the thorax region. We used the attenuation map in the anatomical prior. The results demonstrated that both methods can estimate the misalignment and deform the anatomical image accordingly. However, the performance of the first approach depends highly on the workflow of the alternating process. The second approach shows a faster convergence rate to the correct alignment and is less sensitive to the workflow. The presence of anatomical information improves the convergence rate of misalignment estimation for the second approach but slow it down for the first approach. Both approaches show improved performance in misalignment estimation as the data noise level decreases.