Research On Technique And Application Of A Vision-based Webpage Segmentation

The33rd Statistical Report on Internet Development in China released by CNNIC showed that the number of mobile phone netizen soared from178million to500million people only five years, accounted for81.0percent of all netizens, and maintained a stable growth trend. So the mobile phone netizens have become a large user group in steady growth. However the limitations of mobile phone screen size and computing power caused the problem of mobile browser didn’t properly present and even open web pages designed for PC. With the rapid development of microelectronics and mobile communications technology, the contradiction has become increasingly prominent and sharp. For this problem, VWS technology was proposed to solve this problem that mobile browser can’t accurately and efficiently display web pages, and to improve the user experience.VWS technology identifies the feature of content blocks in a web page from visual perspective, then views webpage segment as grouping optimization problem based on optimization theory. So the webpage is divided into subpages network which composed of semantic complete sub-pages that are suitable for mobile phone display. Last, select a specific sub-page as home page pushing to users, the user can switch and browse each sub-page according to their own needs. The novel technology innovatively puts forward the webpage preprocessing algorithm to combine the webpage content with style information to fully and efficiently use style information. The novel technology innovatively describes visual characteristics of content blocks from six dimensions based on human visual features and defines the similarity calculation formula of content blocks in the visual characteristic aspects. Then, design neural networks to directly confirm the weight of each dimension. The directly determine method is more real and objective than expert experience and more efficient and realistic than the traditional weights determination of neural networks. The novel technology innovatively views webpage segment as grouping optimization problem and designs web page segment algorithm based on the Kruskal algorithm in optimization theory. Under the premise of ensuring the mobile phone display properly, VWS technology can realize the maximize similarity and improve the relevance and integrity of the semantics between content blocks of each sub-page. The high-quality digital learning resources in ECs are as high as1.2564million. With the rapid popularity of informal learning theory and the convenience of mobile phone, more and more learners want to access ECs webpages by mobile browers. So we can apply VWS technology in ECs and take the ECs as experimental subjects to verify the feasibility of VWS technology, to solve the displays problem of ECs webpage on mobile browsers, and to promote the construction and development of ECs. In the experiments, randomly select100different ECs website, and randomly access to a page in each website which may be home page or non-home page. The selected100ECs web pages are segmented by VWS and VIPS technology and the segmentation results are analyzed by the qualitative and quantitative experiments. Analysis results show that VWS technology can be well done Web page segmentation, solve the problem of web pages displayed in the phone rightly and has a better user experience.