Abstract:

The minimum fuzzy entropy thresholding method is commonly used for image segmentation. Based on the dual conception of the fuzzy entropy, the corresponding maximum fuzzy energy image thresholding method is presented. Consider that the threshold obtained by the maximum fuzzy energy threshold method cannot be used to locate the valley of the image histogram in some cases, a revised method is proposed by adding the probability information of the thresholding point. The threshold gotten by the revised method is much close to the valley of the graylevel histogram. Experiment results show that the new segmentation method can get good results for images with an unimodal histogram or with small objects.