Research On Data Representation Of Scalable Vector Graphics

The emerge of SVG(Scalable Vector Graphics) provides a new planar network image format based on XML (extensible Markup Language) feature's of two-dimensional graphics format for Web user. SVG-based data representation is an important content of SVG application, which includes data representation method of SVG entity, SVG structure design and classification, SVG database storage and Software Architecture based on SVG application, etc. Recently, data representation method of SVG entity and SVG classification are main purpose and crucial issue in SVG application research.In this dissertation, research efforts are concentrated on the SVG data representation method, which includes SVG structure design, XML & extensible SVG Documents Classification by Structure and Keywords frequency, Method for SVG-Based Rendering of Raster Images, RDMS storage model based on SVG and Software Architecture based on SVG application. Several effective methods are proposed and validated.For SVG structure design, a vectorlize structure mode and hierarchy method of vector graphics entity is proposed according to vector graphics feature. Then, a description program language based on XML syntax (called XSVGDL) is designed, which can describe hierarchy vector graphic entity and a method is presented, which can be translated XSVGDL document into SVG code because XSVGDL document can't be rendering in Internet browser. It can solve the structuralize problem of SVG document by above all methods. Structure Design of SVG-Based method is good for the content based storing and searching for SVG, and also strengthens the readability and expansibility of SVG code.For XML/XSVG document classification, according to XML/XSVG document character, an efficient method for computing XML/XSVG documents similarity, position and frequency for keywords or vector graphics feature in documents is presented. Then some features are selected from XML/XSVG documents based on our method, and an algorithm of multi-classification of XML/XSVG document is proposed using support vector machines. In our algorithm, we create a CFK (Classifier Feature Kernel) of common similarity features from each sample set of XML/XSVG documents class. The class label of an XML/XSVG document is determined by computing similar distance between test an XML/XSVG document and each CFK. Experimental results prove the effectiveness of our classification algorithm and good performance for multi-classification of XML/XSVG documents.For SVG-based rendering method of raster image, it starts image segmentation for raster image with similar features of pixel; the edges of image segmentation are obtained, merged and compressed. Then the raster image is segmented an image region set which compose of shape feature and color feature. They are rendered easily by SVG with polygon or Bezier curve. Therefore the obtained SVG graphics with different region edges for similar features of pixel are merged if the converted SVG is very large size. The SVG rendering approach is very useful for huge size images and thick texture culture relic images of digital museum.Finally, according to modularized SVG Entity feature and framework, we describe a relation among SVG entity, SVG class and SVG object, and then a data model is presented for SVG application which is storing SVG class or object in RDBS, hierarchical software architecture is designed based on SVG application. Hereby it solves three problems: slow execution speed in SVG application system including very large SVG entities, data reuse issue for SVG component and SVG development issue for based on RDMS. We design an explorer platform system of digital museum based on our method, and compare with traditional digital museum development model based on Html/Image. Experimental results show the effectiveness of our method and good performance for SVG application development.