基于目标图领域点比较的色彩传递研究

Color Transfer Study Based on Target Image Field Point Comparison

目前,图像的色彩传递主要基于Lαβ颜色空间,利用邻域点匹配算法和高阶矩匹配算法,或者采用亮度分割匹配等方法,在源图和目标图之间进行色彩匹配传递。但是,这些算法匹配点数量多,速度慢。针对这个问题,文章在原有匹配方法的基础上提出了基于目标图相邻点匹配以及目标图邻域点匹配的改进方法,在保证传递效果的情况下,提高了色彩传递的速率,文章还讨论了色彩传递过程中存在的一些问题,给出了一些解决方案。

Currently, color transfer between images is based on Lαβ color space,using the neighborhood matching algorithm and higher order moment matching algorithm,or using the brightness segmentation matching algorithm to transfer color between the source image and the target image.However, the number of matching points is large,and the matching rate is slow.For this problem,basing on the original match methods,the article give some improved methods based on the target image＇s adjacent points＇matching method and the target image＇s neighborhood matching method.These methods can improve transferring rate in the case of guaranteed transferring effect,the article also discusses some problems in the color transfer process,then giving some solutions.