提出一种基于变权全变分(RWTV )迭代重建的金属伪影校正算法以抑制CT系统重建图像中出现的金属伪影。该算法应用自定义的权值函数对全变分模型做加权惩罚, 生成权值全变分模型; 通过交替解权值全变分最小化过程和更新权值步骤实现变权全变分重建算法。应用该算法对数值模型和临床图像前向投影生成的投影数据分别进行了重建实验。数值模型实验结果表明: 在60个采样角度下, 用提出算法重建的图像具有最高的空间分辨率特性; 且信噪比值较平滑插值金属伪影校正算法、全变分约束最优化算法的重建结果分别高出17.523 6和 7.145 2 dB。临床数据实验结果表明: 该算法重建结果有效抑制了CT金属伪影, 清晰重建出颅骨内的细节解剖结构, 极大提高了重建图像的质量。

A novel metal artifact reduction algorithm based on a Reweighted Total Variation (RWTV) reconstruction was proposed to suppress the metal artifacts in Computed Tomography (CT) images to improve the image quality. A user-defined weight function was used as penalty weight to generate the weighted Total Variation(TV) model. The reweighted TV reconstruction algorithm was implemented through solving alternately the weighted TV minimization problem and updating the weight procedure. Then, the algorithm was used in the reconstruction experiments for projections from forward projecting phantom and clinical data respectively. The digital phantom experiments show that the proposed algorithm has the topmost spatial resolution of reconstructed images in 60 sampling angles. Furthermore, the signal-to-noise ratios of images through the proposed algorithm are 17.523 6, 7.145 2 dB higher than those of images reconstructed by Smooth Interpolation Metal Artifact Reduction (SI-MAR) algorithm and TV minimization based on Constrained Optimization-Penalized Smoothness (CO-PS) algorithm respectively. Clinical experimental results demonstrate that the proposed algorithm successfully suppresses the metal artifacts and clearly restores the anatomical structure in the skull. Moreover, the quality of reconstructed images has been greatly improved by the proposed metal artifact suppression algorithm.