当前位置:主页 > 管理论文 > 移动网络论文 >

一种基于视觉的网页分割技术及应用研究

发布时间:2018-03-31 20:09

  本文选题:基于视觉 切入点:网页分割 出处:《华中师范大学》2014年硕士论文


【摘要】:CNNIC第33次中国互联网络发展状况统计报告显示,仅5年时间,手机网民由1.78亿猛增至5亿人,占总体网民81.0%,并保持稳定增长趋势,由此可知手机网民已成为稳定增长的庞大用户群体。而手机屏幕尺寸和运算能力的局限性,导致手机浏览器无法正常呈现、甚至无法打开针对PC设计的Web网页。随着微电子技术与移动通信技术的迅速发展,该矛盾日益突出和尖锐。针对该问题,提出VWS技术,以解决手机浏览器无法准确、高效地显示Web网页问题,从而提高用户体验。 VWS技术从视觉角度标识网页中内容块的特征,之后基于最优化理论,把网页分割看作分组最优化问题,将网页分割为语义完整且适合手机显示的子页网。最后,选取特定子页作为首页推送给用户,用户可根据需要切换浏览各子页。该技术创新地提出网页预处理算法,将网页内容与样式信息进行融合,实现样式信息充分、高效地利用;创新地依据人类视觉特征从六个维度描述内容块视觉特征,并定义内容块在视觉特征方面的相似度计算公式,之后设计神经网络直接确定公式中每个维度的权值,此权值直接确定法较专家经验法真实、客观,比传统神经网络权值确定法高效、逼真;创新地将网页分割看作分组最优化问题,并基于最优化理论中的Kruskal算法设计网页分割算法。在确保手机正常显示的前提下,实现各子页中内容块间的相似度最大化,提高了各子页中内容块间的语义的相关性与完整性。 ECs中含优质数字化学习资源高达125.64万,随着非正式学习理论在我国迅速流行加之手机的便捷性,越来越多的学习者希望通过手机访问ECs网页。因此,可以将VWS技术应用于ECs中,并以ECs为实验对象验证VWS技术的可行性,借此解决ECs网页在手机浏览器中的显示问题,增加ECs的访问渠道,从而促进精品课程的建设与发展。实验中随机选取100个不同的ECs网站,在每个网站中随机获取一个网页,采用VWS技术与VIPS技术分割得到的100个ECs网页,并对分割结果进行定性实验与定量实验。分析结果表明,VWS技术可出色地完成Web网页分割,实现针对PC端设计的网页在手机中的正常显示,并且具有较好的用户体验。
[Abstract]:Statistical report CNNIC thirty-third China Internet development shows that only 5 years, mobile phone users jumped from 178 million to 500 million people, accounting for 81% of total Internet users, and to maintain a steady growth trend, so that mobile phone users have become huge user groups to stable growth. And limitations of mobile phone screen size and operation ability, leading mobile phone the browser can not be normal, even unable to open the PC design Web ". With the rapid development of microelectronics technology and mobile communication technology, the contradictions have become increasingly prominent and sharp. Aiming at this problem, put forward VWS technology to solve the mobile phone browser is not accurate, efficient display of Web pages, so as to improve the user experience.
VWS technology from the visual angle of web page content features identification block, then based on optimization theory, the web page segmentation as grouping optimization problem, the semantic web pages into complete and suitable for mobile phone display sub network page. Finally, select a specific sub page as the home page pushed to the user, the user may need to switch the browse each sub pages. Put forward the "technology innovation preprocessing algorithm, web page content and style information fusion, realization of information is full and efficient use of human visual characteristics; according to the creative description of content block visual features from six dimensions, and define the content block in the visual feature similarity formula. After the design of the neural network directly determines each dimension in the formula weights, the weights direct determination method is based on the expert experience real, objective, than the traditional method to determine the weights of the neural network, the innovation, realistic; Web page segmentation as grouping optimization problem and Kruskal algorithm design web page segmentation algorithm based on optimization theory. In the premise of ensuring the normal mobile phone display, the sub page content block similarity between the maximum, improve the relevance and integrity of each sub page content block between the semantic.
Up to 1 million 256 thousand and 400 with high quality digital learning resources in ECs, with the convenience of the informal learning theory in China's rapid and popular mobile phone, more and more learners hope to access the ECs page by mobile phone. Therefore, we can apply VWS technology to the ECs, and with ECs as the experiment on the feasibility as to verify VWS technology, to solve the the ECs page in the mobile phone browser display, increase access to the ECs channel, so as to promote the construction and development of excellent course. 100 different ECs sites were randomly selected in the experiment, random access to a web page at each site, 100 ECs pages using VWS technology and VIPS technology obtained by segmentation, and qualitative experiment the segmentation and quantitative experimental results. The analysis results show that the VWS technology can complete Web page segmentation, design and Implementation for the PC side of the web page in the mobile phone display properly, and has better use Household experience.

【学位授予单位】:华中师范大学
【学位级别】:硕士
【学位授予年份】:2014
【分类号】:TP393.092

【参考文献】

相关期刊论文 前10条

1 王琦,唐世渭,杨冬青,王腾蛟;基于DOM的网页主题信息自动提取[J];计算机研究与发展;2004年10期

2 张雨浓;钟童科;李巍;易称福;;Laguerre正交基前向神经网络及其权值直接确定法[J];暨南大学学报(自然科学版);2008年03期

3 陈翰生;曾剑平;张世永;;一种基于位置信息的Web页面分割方法[J];计算机应用与软件;2009年07期

4 史晶;吴庆波;杨沙洲;;移动终端个性化页面显示优化技术研究[J];计算机工程;2012年18期

5 李军;陈君;王玲芳;倪宏;;一种垂直页面分割与信息提取方法的研究[J];计算机应用研究;2013年03期

6 彭红超;童名文;邹军华;郝秋红;;基于规则的网页分割预处理算法研究[J];计算机科学;2013年S2期

7 王静;姚勇;刘志镜;;基于广义隐马尔可夫模型的网页信息抽取方法[J];山东大学学报(理学版);2007年11期

8 范质彬,王静立,纪震;HTML→WML转码器关键技术的实现[J];深圳大学学报;2002年02期

9 蒙韧;邵延振;袁鼎荣;;一种基于页面Block的Web信息提取方法[J];计算机技术与发展;2010年01期

10 孙晓辉;刘建;王劲林;陈晓;;基于CSS的网页分割算法[J];微计算机应用;2008年09期



本文编号:1692264

资料下载
论文发表

本文链接:https://www.wllwen.com/guanlilunwen/ydhl/1692264.html


Copyright(c)文论论文网All Rights Reserved | 网站地图 |

版权申明:资料由用户d6009***提供,本站仅收录摘要或目录,作者需要删除请E-mail邮箱bigeng88@qq.com