Image Retrieval Based on Color and Texture

Abstract

Color and texture are important image features. In this paper a clustering algorithm based on mean shift is used to extract the dominant color in CIE Lab color space. Earth mover's distance is used to calculate the color dissimilarity. A new method based on texture features is proposed, which calculates the gray-level variation between neighboring pixels at the directions of horizontal, vertical, 45 degree and 135 degree. The technique of relevance feedback is used to enhance the effectiveness. Finally, a prototype system is developed to compare the retrieval precision, the rank and the execution time by two experiments. The results show that the proposed approach is effective.