Abstract: Based on the quotient space and granular computing theory, the image retrieval process was analyzed, and then the method of multi-granularity image retrieval based on quotient was proposed. With the proposed method, an imp ge was divided into different regions based on the equivalent relation R (i. e.，connectivity of image' s dominant color) and features of different regions in different granularity levels of color, shape, spatial information were extracted respec- tively, and then the synthetic feature was obtained by composing the attribute functions in different granularity levels based on the theory of composing multi granularity attribute functions in the quotient space. Finally, the images were re- trieved images by the synthetic feature. Experimental results indicate that the proposed methodin is superior to the method based on single attribute,MI}H and the color volume histogram method.

Key words: Quotient space,Composite attribute function,Granular computing,Image retrieval