一种基于灰关联的PET重建图像评价方法  被引量:1

Quality assessment for PET reconstructed image based on grey relational analysis

在线阅读下载全文

作  者:卢荣辉[1] 陈宗哲 魏晓华[3] 熊孝存[1] 陈荣旺[1] 罗丰 邓秀娟[2] 

机构地区:[1]武夷学院实验室管理中心,福建武夷山354300 [2]百色学院信息工程学院,广西百色533000 [3]武夷学院人文与教师教育学院,福建武夷山354300 [4]武夷山市立医院医学影像科,福建武夷山354300

出  处:《中国医学物理学杂志》2016年第10期1051-1056,共6页Chinese Journal of Medical Physics

基  金:福建省"大学生创新创业训练计划"项目(Sj201210397741);武夷学院青年教师专项科研基金(xq201025);武夷学院校级科研项目(xl201010)

摘  要:提出一种新的基于灰关联分析的图像评价方法,对正电子发射断层成像图像进行质量评价来确定迭代何时停止。计算各种截断频率滤波反投影(FBP)重建图像与数字假体间灰关联度,再用最大灰关联度FBP重建图像作为参考,计算其与不同次数迭代重建图像的灰关联度。实验结果表明,最优的FBP重建图像截断频率为0.3周期/像素,最大似然期望最大化最优图像产生在第10次迭代。灰关联分析与峰值信噪比具有相同的评价结果,研究结果与其他早期研究结论相同,获得最佳正电子发射断层成像图像所需的迭代次数与临床相同。An image quality assessment method for positron emission tomography (PET) image based on grey relational analysis is proposed in the paper to assess the PET image quality and determine the stopping point of iteration. The grey relational grades between the images with different cutoff frequencies constructed by filtered back projection (FBP) and the digital phantom are calculated. Taking the FBP reconstructed image with maximum grey relational grade as the reference, the grey relational grades between the reference and the reconstructed images with different number of iteration are calculated. Experimental results show that the optimal cutoff frequency of FBP reconstructed image is 0.3 cycle/pixel, and that the image with best quality of maximum-likelihood expectation maximization is found at the 10th iteration. The results of grey relational analysis are consistent with those of peak signal noise ratio, and the research results are also consistent with the conclusions of previous studies, obtaining the number of iteration needed for optimal PET image which is the same with daily clinical use.

关 键 词:灰关联分析 正电子发射断层成像 峰值信噪比 质量评价 最大似然期望最大化 

分 类 号:R445[医药卫生—影像医学与核医学] N941.5[医药卫生—诊断学]

 

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