An automatic parameter selection for fuzzy entropic optimal threshold to enhance the stumpy foreground objects in the images based on a fuzziness measure is presented in this paper. This work is improvement of an existing method. Using fuzzy logic, the problems involved in finding the minimum of a criterion function are avoided. Similarity between gray levels is the key to find optimal threshold to initialize the regions of gray levels which are located at the boundaries of the ROI and after that by using an index of fuzziness a similarity process is started to find the thresholding point. The advantages of the proposed method over their conventional counterparts are fourfold. First, the automatic fuzzy parameters for optimal threshold are selected. Second, the ROI is used instead of the whole image, so any irregularity outside the ROI will have no influence on estimating the threshold. Third, it provides a mechanism to handle cost differences of different types of classification error in response to practical requirements. Fourth, is by appropriately specifying the lower a upper bounds of the background proportion within the constrained gray level range, the proposed method yields substantially more robust and more reliable segmentation.