Speeding the vector search algorithm for regional color channel features based indexing and retrieval systems

Color indexing is a technique by which images in the database could be retrieved on the bases of their color content. In this paper we propose a new algorithm for matching an input query with the color database images. The proposed algorithm is applied on the similarity retrieval system which was proposed in [Appas 99] to show the effect on the retrieval speed. The proposed algorithm adds one feature to an image feature vector. The added feature is obtained from the color channel features of the image regions. Comparing query hit percentage of image retrieval of the proposed algorithm in [Appas 99] with the new matching algorithm and with the normal exhausted matching algorithm shows that the performance of the retrieval system is nearly the same in both of cases but the retrieval speed increases specially with the large size databases when using the proposed matching algorithm.