新的Jensen-Schur测度在医学图像配准中的应用  

Novel measures based on Jensen-schur for medical image registration

作　　者：胡顺波[1] 刘常春[2] 杨金宝[2] 

机构地区：[1]临沂师范学院物理系,山东临沂276005 [2]山东大学控制科学与工程学院,山东济南250061

出　　处：《光电子．激光》2008年第10期1419-1422,共4页Journal of Optoelectronics·Laser

基　　金：国家"863"计划资助项目(2006AA02Z4D9);山东省自然科学基金资助项目(Z2006C05)

摘　　要：针对传统的互信息和归一化互信息运算时间长、收敛性能和抗噪声能力差的缺点,提出了一种新的医学图像配准测度,Jensen-Schur测度。从理论上分析了该测度可用于图像配准的性质,并构造了三类Jensen-Schur测度:Jensen-Schur-alpha、Jensen-Rényi-alpha、Jensen-Schur-w测度。通过对人体脑部MR/CT和MR/PET图像的刚体配准实验,从运算时间、收敛性能、抗噪鲁棒性方面,对这三类测度、互信息和归一化互信息进行了比较和分析。实验结果表明,Jensen-Schur-alpha(alpha=2或3)和Jensen-Schur-w测度的收敛性能优于其它测度,运算速度快于其它测度,对噪声有很强的鲁棒性。最后用Jensen-Schur-alpha(alpha=2)测度进行多模态医学图像的配准实验,结果表明效果良好。In order to reduce the registration time, improve the convergence ability and perform better in the presence of noise, a novel medical image registration measure is proposed, which is called Jensen-Schur measure. According to the Jensen-Schur measure, three kinds of measures are constructed： Jensen-Schur- alpha.Jensen-Rényi-alpha.Jensen-Schur-w. The characters of three measures,mutual information and normalized mutual information are analyzed and compared by applying them to rigid registration of magnetic resonance （MR）/computed tomography （CT） and MR / positron emission tomography （PET） images. The results show that Jensen- Schur- alpha （alpha= 2 or 3） and Jensen-Schur-w measures outperform other measures in the computational speed, convergence performance and noise immunity. A multi-modality medical image registration test is conducted and the good result is obtained.

关 键 词：图像配准 Jensen-Schur测度 互信息 

分 类 号：R318.51[医药卫生—生物医学工程]