Multi-granular method for retrieving Thangka images

Based on quotient space and granular computing theory, the image retrieval process is analyzed, and a multi-granular method based on quotient space is proposed for Thangka image retrieval. With the proposed method, an image is divided into different granularities based on hierarchical thinking, and different quotient spaces are constituted according to different behaviors under different granularities of the image. The synthetic features are obtained by composing the attribute functions at different granularity levels based on the theory of composing multi-granularity attribute functions in the quotient space. Finally, the images are retrieved by the synthetic method. The experiment showed that the synthetic method could be able to effectively retrieve Thangka images.