Using a fuzzy morphological structural element for image segmentation

Management of uncertainty is an important problem in the development of automated systems for scene interpretation, particularly in image segmentation algorithms. In this paper a fuzzy-based morphological structural element is introduced to be used in image segmentation based on pixel classification, with the aim to reduce uncertainty problems. The shape, size and structure of this type of element have been defined in order to reduce pixel misclassifications providing boundaries among regions as accurately as possible.