Operation Of Granular Computing With Binary Relation And Its Application In Image Retrieval

Granular computing has been widely used in data mining, knowledge discovery,artificial intelligence and other fields as a popular intelligent information processing technology. With the increasingly complex data of practical problems in application,how to build information granules operational rules,it's important to effectively solve the question by utilize granular computing method.For this problem, this paper makes two works, one is analyzed the operational rules of information granules with binary relation, and the second is apply the operational rules of information granules to the field of color card image retrieval.The main research work and achievements are as follows:1. This paper gives some related properties on the basis of previous work of our research team,that the related operations of information granular,.Based on the research result of granular computing with nonstandard analysis by our research team on preliminary, this paper defines the difference and inverses operational rules of information granules, so that the operation theory of information granules on granular space can be extended to provide better implement technology for the mutual transformation, synthesis and decomposition of information granules inter layer on granular space. The related work above provides a theoretical guidance for the further mining of the mutual transformation relationship of different information granules inter layer in the same granular space.2. Apply the operational rules of information granules with the equivalence relation or neighborhood relation researched above to test the feasibility and effectiveness of the theory and method researched.Firstly, it proposes the multi-granularity color features partition method based on image color features combines with granularity thought, and the muti-granularity textural structure method based on different dimension histogram of LBP texture.Then proposes the image granular space based on various information granules according to different image features, and designs the image retrieval example to analyses the difference and inverse operational rules of information granules with thebinary relation based on the image granular space.3. Propose image retrieval model based on granular computing on granular space based on the image granular space. It proposes the detail model frame and design the image retrieval algorithms with equivalence relation information granules computing and neighborhood relation information granules computing.Firstly, the image retrieval model proposed in the article has been tested and analyzed on the real image data sets acquired from bags enterprises. Secondly, test and analyses the image retrieval method based on various granular feature and information granules computing proposed in this article with image retrieval methods based on image features,based on classical LBP texture feature and based on color features and texture features. The test result shows the image model proposed in this paper is effective.Finally, give the summary and outlook on the research of this paper.