Structure- and Content-Based Retrieval for XML Documents

Abstract

The XML was proposed as a standard markup language to make Web documents in 1996 (Extensible Markup Language, 2000). It has as good an expressive power as SGML and is easy to use like HTML. Recently, it has been common for users to acquire through the Web a variety of multimedia documents written by XML. Meanwhile, because the number of XML documents is dramatically increasing, it is difficult to reach a specific XML document required by users. Moreover, an XML document not only has a logical and hierarchical structure in common, but also contains its multimedia data, such as image and video. Thus, it is necessary to retrieve XML documents based on both document structure and image content. For supporting the structure-based retrieval, it is necessary to design four efficient index structures, that is, keyword, structure, element, and attribute index, by indexing XML documents using a basic element unit. For supporting the content-based retrieval, it is necessary to design a high-dimensional index structure so as to store and retrieve both color and shape feature vectors efficiently.