二维图像强度-几何相关系数的统计研究方法

Statistical Method of Intensity-Geometrical Correlation Coefficient for 2D Images

将Vizilter的图像强度-几何相关系数的研究方法加以推广,对随机马赛克图像,采用均方范数作为量化方法,研究了基于投影算子的中心化形态相关系数和几何相关系数更一般的形式,并以Vizilter的各种相关系数作为其特例,研究结果更符合图像的自然构成.此外,通过引入序化函数,研究了相关系数的一些新形式,使得图像相似度比较指标更加丰富.

In this paper,the research method of Vizilter′s image intensity-geometric correlation coefficient is generalized.The mean square norm is used as a quantitative method for random mosaic images.The central morphologic correlation coefficient and the geometric correlation coefficient based on the projection operator are researched more generally.And the various correlation coefficients of Vizilter will be used as a special case.The result of the research is more consistent with the natural composition of the image.In addition,by introducing the ordering function,some new forms of the correlation coefficient are researched,which makes the image similarity comparison coefficient index more abundant.