摘要
为消除CT检查过程中病人运动引起的伪影,提高图像质量,提出一种基于HLCC多约束运动参数估计的校正算法.首先用贝塞尔曲线模型描述病人的刚性平移运动,基于HLCC构造一个包含运动信息的目标函数,再引入运动幅度的有限性和图像零阶几何动量为常量作为约束条件,通过二次规划的方法估计运动参数;重建过程中改进滤波反投影算法,加入运动参数进行补偿,以消除运动伪影.实验结果表明,该算法能够准确地估计出物体的运动参数,有效消除运动伪影,具有较强的鲁棒性.
To eliminate artifact caused by patient motion during a CT scan, an algorithm based on HLCC was proposed to estimate the parameters of translational motion. The translational motion as Bezier curves was modelled, an objective function containing the motion information based on HLCC was developed and limits of translation excursion and zero order moment of the object as constraints were introduced. The motion parameters were estimated by solving the optimization problem with quadratic programming method. The estimated motion parameters were introduced into an improved filtered back-projection reconstruction algorithm to correct motion artifacts. Experimental results show demonstrate the accuracy and robustness of the proposed method.
