基于空间自适应全变分的图像配准  

Image Registration Based on Spatial Adaptive Total Variation

作　　者：张群艺 杨奋林[1] ZHANG Qunyi;YANG Fenlin(School of Mathematics and Statistics, Jishou University, Jishou 416000, Hunan China)

机构地区：[1]吉首大学数学与统计学院,湖南吉首416000

出　　处：《吉首大学学报（自然科学版）》2020年第5期37-40,共4页Journal of Jishou University(Natural Sciences Edition)

基　　金：国家自然科学基金资助项目(11501243);湘西自治州科技计划项目(2018SF5021)。

摘　　要：通过引入空间自适应正则化参数,建立了离散的空间自适应全变分配准模型,并结合一阶泰勒展开式给出了有效的不动点求解方法.实验结果表明,由空间自适应全变分配准模型得到的图像的质量比全变分配准模型的更好,迭代时间更短.A discrete spatially adaptive total variation image registration model is established by introducing spatial adaptive regularization parameters,and an effective fixed point method is given by combining the first-order Taylor expansion.Experimental results show that the image obtained from the spatial adaptive total variation registration model has better quality and requires less iterative time than the total variation registration model.

关 键 词：图像配准 空间自适应 正则化参数 不动点方法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]