A novel image retrieval algorithm for computer aided color design

At any time in the process of product color design, designers need to initialize a color design scheme and then use it to retrieve existing similar designs from related products, for referencing and stimulating the ideas of color design. Previous researches on color image retrieval consider the color of an image as a whole eigenvector for efficient computation of color similarity. But they are not suitable for a computer aided color design tool development. In this paper, we propose a framework of a computer aided color design tool and a novel image retrieval algorithm based on a layered HSV model to calculate H, S and V histogram similarity separately and apply them sequentially to retrieve related color designs gradually. The experimental results show that the proposed color design tool can assist designers to effectively retrieve similar color design images for referencing based on a user specified color design scheme.