磁共振图像处理中部分傅里叶重建算法的比较  被引量:3

Comparison of partial Fourier reconstruction algorithms in MRI image processing

在线阅读下载全文

作  者:何汶静[1] 陈晓文[1] 朱高杰 罗海 

机构地区:[1]川北医学院医学影像学系,四川南充637000 [2]奥泰医疗系统有限公司,成都610000

出  处:《重庆医学》2016年第20期2804-2806,2809,共4页Chongqing medicine

基  金:四川省教育厅资助课题(1BZB0237)

摘  要:目的比较磁共振(MRI)图像处理中部分傅里叶重建算法,对比各算法的优缺点,改进算法。方法基于部分傅里叶重建算法理论上的k空间数据的共轭、对称性,利用1.5T医用MRI设备,采集全k空间的数据,重建时仅利用其中55%~85%的数据模拟部分采集方式,用零填充、共轭对称、Homodyne、POCS几种算法分别重建图像。结果共轭填充方式振铃伪影最明显,POCS算法去振铃伪影效果最好。结论POCS算法多次用低频相位加权迭代,在去掉伪影的同时造成高频图像细节丢失,且耗时较长;Homodyne算法能在保留图像细节和消除振铃中间取得较好的平衡,如果用Sinc函数作为权重函数进行相位解缠绕,还可以进一步提高图像质量。Objective To compare the various Partial Fourier reconstruction algorithms in MRI image processing and their advantages and disadvantages for improving algorithm. Methods Based on the conjugate symmetry in k-space data in the partial Fourier reconstruction algorithm theory, the 1.5T medical magnetic resonance equipment was utilized, the k-space data were collected,in reconstruction,only 55%-80% of full data were utilized to simulate the partial collection pattern and reconstruct the images by using the several algorithms of zero filling,conjugate synthesis,Homodyne detection and POCS. Results The results presented that the ringing artifact was most obvious in the conjugate filling pattern and the POCS algorithm had the best effect for eliminating the ringing artifact. Conclusion The POCS algorithm repeatedly uses the weighted iteration with low frequency phase positions, which eliminates artifact,meanwhile causes the lose of high frequency image details, moreover the time consuming is longer ;the Homodyne algorithm can keep a good balance between maintaining the image details and diminishing ringing. If the Sinc -function serves as a weighting function for conducting phase unwrapping,the image quality may be further improved.

关 键 词:部分傅里叶重建 磁共振成像 K空间 

分 类 号:R445.2[医药卫生—影像医学与核医学]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象