期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2014
卷号：3
期号：11
页码：17567
DOI：10.15680/IJIRSET.2014.0311067
出版社：S&S Publications
摘要：In this paper, a new accurate su bpixel regis tration for puresh ift es timation is pro posed. The noise effect, which disturbs the quality of regis tration proces s , is taken into account. The kern el reg ress ion method which represents the fieldofnon para metric statistics is used as atool for the estimation process dueto it spowerfu l capabilities in the field of both denoising and interpolation . The kernel regression depends on studyin galocal region inten sities distribution and gradients. By ap plying gradient descentmethod, the global translation parameters can bees timated. Exper imen tal results show that our propos ed method can estimate the tran slation parameters accu rately. Fu rthermore, our method performs well in noisy images.
