Weighted fuzzy correlation for similarity measure of color-histograms

In this report, fuzzy correlation of color-histograms is proposed for the similarity measure between the color contents of images. By employing a weighted algorithm on the process of calculating the peak value of the correlation, it offers the flexibility to adapt to an image and its best matching images with partial color distortion or local color characters. Experimental examples demonstrate the effectiveness of the proposed method in its application to the retrieval of color images of interest specified by their color content.