基于视觉特征的网页正文提取方法研究  

The research on vision-based Web page information extraction algorithm

作　　者：安增文[1] 徐杰锋[1] 

机构地区：[1]中国石油大学(华东)计算机与通信工程学院,山东东营257000

出　　处：《微型机与应用》2010年第3期38-41,共4页Microcomputer & Its Applications

摘　　要：利用网页的视觉特征和DOM树的结构特性对网页进行分块,并采用逐层分块逐层删减的方法将与正文无关的噪音块删除,从而得到正文块。对得到的正文块运用VIPS算法得到完整的语义块,最后在语义块的基础上提取正文内容。试验表明,这种方法是切实可行的。To get the useful information blocks, this paper first segmented the Web page into blocks with its visual features and its DOM tree＇s characteristics, and then deleted the noise blocks. This is a recursive process until no block can be deleted. Then handled the reserved blocks with the VIPS algorithm to get the semantic blocks. At last, got the text content by handling the semantic blocks. Experiment shows that this method is feasible.

关 键 词：页面分块 信息提取 视觉特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]