Font Size: a A A

A Look Into The Image Semantic Marking Strategy Based On Open Logic

Posted on:2011-12-19Degree:MasterType:Thesis
Country:ChinaCandidate:T LinFull Text:PDF
GTID:2155360302497531Subject:Logic
Abstract/Summary:PDF Full Text Request
The stockpiling of online image data has forced us to develop equivalent countermeasures for the task of data management. The easiest and most convenient way of issuing an online search is to input desiring semantic concepts via text data into a web based interactive interface, yet so far, no satisfying solutions of automatic content matching were provided on semantic level.The principal means available now can be divided into TBIR, Text Based Image Retrieval and CBIR, Content Based Image Retrieval. TBIR is related to image semantics for the background info given by image context, CBIR, somehow, is semantic irrelevant unless particular label(s) were assigned to that image character. But none of them would be feasible for a real-time semantic search engine applied onto the ever extending Internet and its massive data. TBIR extracts only limited information from the context that will hinder the accuracy of the search outcome, while CBIR deals simply the similarity between given image features, demanding an accurate pairing strategy for complex semantic info and its image feature counterpart that only practical in theory.Clearly we cannot rely on TBIR or CBIR alone for their technical limitations, nor the conjoining of the two. Observing that CBIR is data friendly for its essence of being an image analysis tool and TBIR is user friendly on that it serves as an approximation to human expression, hence here in this article, we proposed the insertion of a third layer:the middle layer, a semantic network alike deduction system, then introduce open logic (Li Wei) as the maintenance of this massive knowledge base of deduction rules and facts. Through this layer, the framework of image semantics was outlined, and the links between image feature and advanced semantics were established. If CBIR can ever be compared to the primal processing of human vision, and TBIR to be compared with the language expression, then our system is unique on that it acts like the human brain, and the human brain, as we all know, is where secrets hidden.
Keywords/Search Tags:Image Semantics, CBIR, Open Logic, AI
PDF Full Text Request
Related items