Fuzzy segmentation-an important tool in image processing

Describes a study of two fuzzy based segmentation algorithms to model successfully ambiguities and uncertainties inherent in assigning an image pixel to an object of interest and to help choosing the threshold for binarizing images using the index of fuzziness and fuzzy compactness. The methods detect object contours to different degrees of certainty on the basis of the information provided by the user, who define a reference point inside the object around which is construct the contours. First, we obtain a connectedness or a similarity map related to the indicated reference point and then for the pixels surrounding that point, if their membership-values are high they are regarded as being inside the object. We also present the connectedness and similarity maps as a useful tool in image processing, so as to obtain better gradient images