摘要
在MR扫描过程中患者的自主或非自主运动使获得的采集数据的相位发生偏移,因而重建图像含有运动伪影,严重地影响临床诊断。由于平移和旋转是最基本的刚性运动,故几乎所有刚性运动都可由他们复合而成。本研究针对MR图像刚性平移运动伪影提出了一个新的后处理方法:逆向迭代修正法(IIC)。通过逆向模拟患者的平移运动,逐次修正信号在K空间中已偏移的相位,然后根据最小熵约束准则,判断所模拟的逆向运动平移量是否等于患者的实际运动平移量,即修正到原来的位置。实验表明,应用本研究提出的逆向迭代修正算法,运动伪影能被修正到一个较为理想的状态。在修正后图像质量及计算耗时方面,优于其他算法。
In the process of MR scan, because patient's motion affects the collected data, the reconstructed image would contain severe motion artifacts. Translation and rotation are the most basic rigid motion, and nearly all of complex rigid motions consist of them. In this paper, a new post-processing method named inverse iterative correction (IIC) is presented based on the principle of generating MRI rigid translational motion artifacts. Through inversely simulating the patient's translational motion, IIC algotithm can be iteratively implemented to correct the damaged phases in K-space signals. According to the criterion of minimum entropy in the imaging field, we can judge if the translational distance of the simulated inverse motion equals that of the patient's actual motion, i.e., being corrected to the original position. The experiments show that through inverse iterative correction, the movement artifacts could be modified to a comparatively ideal state when the entropy is converged to the minimum value of the whole.
出处
《中国生物医学工程学报》
EI
CAS
CSCD
北大核心
2004年第4期353-358,共6页
Chinese Journal of Biomedical Engineering
基金
国家自然科学基金重点资助项目(30130180)
广东省自然科学基金重点资助项目(010583)
关键词
运动伪影
后处理
逆向迭代修正(IIC)
Algorithms
Computer simulation
Data acquisition
Equations of motion
Image reconstruction
Iterative methods
Motion estimation
Patient monitoring
