基于HEA算法的二维最大相关准则图像分割

Image segmentation based on two-dimensional MCC and HEA

针对目前一些算法对于同一图像每次分割得出的阈值不同,存在分割不稳定的问题,提出了一种基于和声五元素算法(HEA)的二维最大相关准则图像分割算法。该算法利用图像目标和背景分布的相关量最大值来选取阈值,并运用HEA算法优秀的全局搜索能力优化运算速度,提高图像分割的稳定性和质量。实验结果表明本文方法有较好的稳定性,分割效果良好。

A two-dimensional Maximum Correlation Criterion（MCC） image segmentation based on Harmony Element string Algorithm（HEA） is proposed to resolve the instability of image segmentation algorithms. In this algorithm, the segmentation threshold is selected by using the maximum correlation value of the distribution of the object and background. The stability is optimized by using HEA which has been proved to be an excellent global search method. Experiment results show that the proposed algorithm can improve the stability and quality of image segmentation.