二维属性直方图的Fisher准则图像分割及快速递推算法

Image Segmentation Based on Two-Dimensional Bound Histogram Fisher Criterion and Its Fast Recursive Algorithm

利用图像中目标和背景之间类间方差和类内方差在类别分离性中的作用,提出了基于二维属性直方图的Fisher准则分割方法.首先,在考虑图像中心像素与邻域中非直接相邻像素的基础上,通过图像直方图的统计分布特性构造属性集,建立新的二维属性直方图.然后根据最大化Fisher准则,获取最优二维阈值向量.同时为降低二维阈值算法的复杂性,提出了快速递推算法.该快速递推算法中,将二维Fisher准则的计算写成递推的形式,减少了大量的重复计算.实验结果表明,所提出的方法不仅能得到理想的分割结果,而且计算量大大减少,达到了快速分割的目的.

View the effect of between-class （between object and background） scatter and within-class scatter in the image simultaneously in the field of ability of classification, image thresholding segmentation method based on 2-D （dimensional） bound histogram and Fisher criterion is proposed. First, the bound set of an image and its corresponding two-dimension bound histogram is constructed considering the center pixel and its neighbor pixels. Then using the maximizing the Fisher criterion, the 2-D threshold vector is obtained, and in order to reduce the computation complexity, a fast recursive algorithm is presented. In the fast recursive algorithm, the computation of 2-D.Fisher criterion is written in recursive form, so that many repeated calculations are avoided. Experimental results show that the proposed method can not only obtain ideal segmentation results but also decrease the computation cost reasonably, and it is suitable for real time application.