A New Method for Gray Level Image Thresholding Using Spatial Correlation Features and Ultrafuzzy Measure

Abstract

One of the most recent techniques employed to estimate an optimal threshold of a gray level image for segmentation is ultrafuzzy measures. In this paper, we introduce relative fuzzy membership degree (RFMD) taking spatial correlation among the pixels in the image into account. We also propose a novel thresholding technique by combining two-dimensional histogram, which was determined by using the gray value of the pixels and the local average gray value of the pixels using ultrafuzziness and RFMD. Compared to fuzzy membership degree, RFMD of type-II fuzzy sets and ultrafuzzy measure is able to better segment critical gray level images. It was observed that the outcome is so encouraging in objective and subjective perspectives over the existing method for all varieties of images.