Binding Semantics Annotation To Visible Multimedia Data

With the rapid development of computer technology and network technology, ourwork life, study life, leisure life is changing fast:(1) Multimedia data get hugersignificance for each person;(2) The form of multimedia data is more and morediversiform;(3) The use of multimedia data is more and more frequently;(4) Theamount of multimedia data become more and more large. However, it is more andmore difficult to extract t useful information from larger, more complex multimediadata, real information is still relatively scarce. These make the management ofmultimedia data become more and more important.Multimedia information contains some very complex semantic feature, thesemantic feature is often a sense of people, is a kind of subjective emotion, it isdifficult to be described. At the same time, multimedia information contained ndifferent forms, such as text, image and video, and has semantic relation with eachother, semantic annotation allows people to make effective use of multimediainformation resources, it is also an important way to access multimedia informationfast and comprehensively. This technique is more and more in urgent need of people,is of great significance to study and research.This article tell about semantic annotation form three aspects of multimedia data(text, image, and video), by using a ontology model example of some domain to solvethe following problems:(1) What features can be extracted from the multimedia dataand how to extract;(2) How to assign weights among these characteristics in semanticannotation;(3) How to make different multimedia data get semantic association. Thisarticle is aimed to improve the efficiency of management, by solving these threeproblems, The method can make better use of multimedia resources applied to manyfields.