A Fuzzy Similarity Measure of Intuitionistic Fuzzy Sets for Color Image Retrieval Systems
Fatemeh Afsari, Esfandiar Eslami and Peng-Yung Woo
In this paper a new scheme for image retrieval is proposed that represents color images by using the concepts of intuitionistic fuzzy sets while the concept of similarity is also characterized by fuzzy concepts. Color features that could be expressed in various color representation systems have been intensively used, independently or jointly, in image processing during the past decades. Fuzziness arises naturally from the imprecision or vagueness of the pixel color values and human perception. The Hue and Value in the HSV color space are used to construct the 2-D intuitionistic fuzzy sets. In the intuitionistic fuzzy sets, not only the membership degrees are considered but also the uncertainties that are involved in the non-membership degrees and are known as the hesitation measures are considered. The proposed similarity measure of intuitionistic fuzzy sets is a fuzzy quantity, rather than crisp quantity due to the vague concept of the similarities. Large-scale experiments demonstrate the robustness and effectiveness of the proposed scheme.
Keywords: Image processing, Intuitionistic fuzzy sets, Image retrieval, Fuzzy similarity measure