A new image enhancement method Type-2 Possibilistic C-Mean Approach

Abstract

Images and visual understandings are basis in everyday life and are very important tool for decision making. However, for improving the image appearance to a human viewer, or to convert an image to a format better suited to machine processing, enhancing methods should be used. There are wide varieties of techniques for this purpose including, contrast and histogram modification, de-noising, statistical methods, and clustering. Among these techniques, clustering especially fuzzy clustering methods are among the most efficient methods that classifies each data into more than one cluster. In the literature, many fuzzy clustering methods have been presented such as Fuzzy C-Mean (FCM) and Possibilistic C-Mean (PCM), which uses Type-1 fuzzy logic. However, Type-2 fuzzy logic can provide better performance, especially when many uncertainties are presented. In this paper, we applied Type-2 fuzzy clustering method for enhancing the images and proposed a new fuzzy Type-2 Possibilistic c-means clustering (PCM) method. The performance of the proposed method in having good results is evaluated by using 6 images.