ITK Metric在卷烟小包装图像配准中的性能比较  

Performance Comparison of ITK Metric in Cigarette Case Image Registration

作　　者：朱梅华 魏承科[2] 耿安鹏[2] 张素伟[3] 

机构地区：[1]广西壮族自治区烟草专卖局,广西南宁530022 [2]华北计算技术研究所,北京100083 [3]太极计算机股份有限公司,北京100083

出　　处：《计算机与现代化》2013年第6期90-94,共5页Computer and Modernization

摘　　要：为了验证医学图像处理框架ITK在卷烟小包装图像配准中的可用性及其性能,通过分析ITK配准框架的原理及配准流程,采用固定图像变换算法、线性插值算法和优化算法,变换相似度测度算法的方式,比较ITK中不同的相似度测度算法在针对中华硬盒卷烟小包装的图像配准中的效率,最后通过大量的实验,将此框架用于不同卷烟小包装的图像配准中,验证了ITK在非医学领域的可用性,得出一种对不同卷烟小包装的图像配准具有一定普适性的配准流程。In order to verify the usability and performance of the medical image processing framework ITK in the image registration of cigarette case, through the analysis of the principle of the IrK registration framework and registration process, using a fixed image transform algorithm, linear interpolation algorithm and optimization algorithm, the similarity measurement algorithm is transformed, the efficiency of the image registration in Zhonghua hard box cigarette packaging is compared by using different similarity measure algorithm in ITK. Finally, through a lot of experiments, the usability of the ITK in the medical field is verified, a certain universal image registration process of different small package of cigarette is draw.

关 键 词：ITK 图像配准 相似度测度 卷烟小包装 

分 类 号：TP391.5[自动化与计算机技术—计算机应用技术]