IMAGE THRESHOLDING USING FUZZY CORRELATION CRITERION AND HARMONY SEARCH ALGORITHM

摘要

This paper reports a novel image thresholding method based on fuzzy set theory and maximum correlation criterion using harmony search algorithm. In this study, the maximum fuzzy correlation criterion is defined using Z- and S-fuzzy member function on image gray level histogram. Then fuzzy correlation criterion image segmentation based on harmony search algorithm is implemented. The experimental studies were conducted on a variety of images by testing the proposed method and some classical thresholding methods. The experimental results demonstrate that the proposed method can select the threshold automatically and effectively. Compared with the exhaustive search method, the harmony search algorithm can give high degree of accuracy while needing less search time and has good search stability in the segmentation experiments.