The Research On Methods Of Web-oriented Multimedia Information Extracting And Its Application In Education

With the rapid development of Internet, there are a lot of multimedia information resources in the Web. It becomes more and more difficult to look for interesting resource because of disorder, distribution and explosive increase of multimedia information resources in the Web. Constructivism puts emphasis on using information resources to support the learning and thinks that constructivism learning environment should offer relative information resources which are selective, abandunt, convenient for learners including image, audio, video, flash and so on. Web multimedia information resources play great role in construction of the constructivism learning environment and learning of the learners. Therefore, this thesis researches the method of web-oriented multimedia information extracting and its application in elementary education-oriented individuality system of indexing web multimedia. This thesis research impacts construction of the constructivism learning environment actively.The thesis summarizes the present methods of extracting information of web multimedia, separates them into two categories: the method based on multimedia content and the method based on text, and points out their merits and faults. The method of extracting multimedia information based on text is researched and realized.Web multimedia information resources often embed in Web pages with some relevant describing text. Multimedia relevant text that embeds in Web page and has semantic relation with multimedia includes the link text, the anchor text, the surrounding text and the environment text. The thesis which based on the structure and revelatory rules to web page's segmentation and the area semantic identification, realizes the understanding of web page's semantics, and presents the definition of the Web multimedia relevant text. Besides that, combining with its distribution characteristics, it has adopted three levels analysis method to carry on the extraction, including the individual level, the area level and the page level, thus realized the Web multimedia relevant text accurately to extract. In order to extract semantic information of multimedia, semantic dictionaries such as topic dictionary, main body dictionary and attribute dictionary are built. And the arithmetic of extracting them is brought forward. Besides, the arithmetic of automatic segmenting words based on the dictionary is presented.Based on the work, the elementary education-oriented individuality system of indexing web multimedia is designed. The system is divided into six parts: the part of individual meta search engine, the part of content analysis, the part of extracting meta information, the part of analysis of web page's structure and content, the part of extracting texts and semantic information and the part of saving multimedia information. Integral structure, database structure, key program technique and topic words selection of different classes and subjects are introduced in detail. The experiment is made, and the result proves that the method is effective. In last, the thesis brings forward the ways that need to be researched more deeply.