Merged-color histogram for color image retrieval

In conventional histogram-based image retrieval algorithms, they usually only find the intersection areas of the color-component distributions of images and thus work well in matching images with exact colors instead of similar colors, especially for computer generated pictures. This could be greatly affected by the overall variation, such as intensity change. In this paper, a novel merged-color histogram (MCH) method for color image retrieval is proposed to facilitate the matching between similar colors by means of color quantization and palette merging. In color quantization, it compacts the color information and matches each color instead of color components, and thus matching of similar colors are accomplished by using palette merging. Experimental results show that the proposed MCH method achieves about 11-32% more precise in the first 20 retrievals for the same image query, and is able to recall 14%-23% more relevant images in the first 100 retrievals, as compared to the conventional RGB-based histogram method.