无约束最优化图像重建算法研究  

Image Reconstruction Algorithm Based on Optimization without Constraint

作　　者：惠苗[1] 潘晋孝[1] 

机构地区：[1]中北大学信息与通信工程学院,山西太原030051

出　　处：《中北大学学报（自然科学版）》2007年第3期256-260,共5页Journal of North University of China(Natural Science Edition)

基　　金：山西省自然科学基金资助项目(20051042)

摘　　要：在无约束最优化为基础的图像重建问题中,为了提高迭代效率以及重建图像质量,首次提出将变度量法应用到图像重建中.分别给出了变度量法,最速下降法和共轭梯度法的算法描述,并且分析了这些方法的优缺点.变度量法在迭代效率和重建图像质量方面优于其它两种算法.通过模拟实验,分别对这三种算法进行了迭代次数和重建图像质量的比较,结果证明了变度量法的优越性.In the image reconstruction based on optimization without constraint, the variable metric method, the steepest descent method, and conjugate gradient method were applied to image reconstruction to improve iterative efficiency and reconstructed quality, and their virtue and shortcoming were analyzed. Variable metric method was better than other algorithms in iterative efficiency and reconstructed quality. Simulating experiment for above algorithms was compared at iterative efficiency and reconstructed quality respectively. The result showed the advantage of variable metric method.

关 键 词：图像重建 变度量法 最速下降法 共轭梯度法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]