An Efficient Similarity Measure for Color-Based Image Retrieval

Similarity measures are an important factor in the Content-Based Image Retrieval (CBIR). This paper finds the most efficient similarity measure from four image similarity measures. Related work on (CBIR) indicated that these measures have significantly improved the retrieval performance. These measures are the Chi-Squared, The Weighted Mean Variance distance (WMV), The Euclidean distance, and Cosine distance. A sample of 50 colored images is selected from CALTECH visual database. These images were transformed to (HSV) color space. Color features were extracted; these features are the color moments. Experimental results of the proposed work show that the Euclidean distance measure is the most efficient measure for color based image retrieval. 1Introduction Recently, a new application field is born via the amount of visual information. Content-Based Image Retrieval has become an active research area. The reason for this is the fact that world wide networking allows us to communicate, share, and learn information in the global manner. Digital library and multimedia databases are rapidly increasing. Therefore, efficient search algorithms need to be developed based on comparison operation and image indexing. Images would be indexed by their own visual contents, such as color, texture and shape so that,