基于HTML文档的WEB图像分类研究  

Research on classification of image on the Web based on Html

作　　者：杨俊梅[1] 张文进[1] 牛又奇[1] 

机构地区：[1]南京农业大学信息科学技术学院,江苏南京210095

出　　处：《农业网络信息》2006年第11期19-22,共4页Agriculture Network Information

摘　　要：随着信息技术的快速发展和Internet的普及应用I,nternet已经成为包含各种信息形式,如文本、图像、视频、音频等的巨大信息资源库。然而由于Web图像从主题、格式,到功能和来源都多种多样,很难用一个统一的标准进行分类。本研究首先从Web页面上图像的功能角度对图像进行初步分类,将其分为链接类、装饰类、广告类和内容类。初步分类可提高Web图像处理的效率和性能,并为它基于Web的应用与研究提供帮助。With the fast development of information technology and broad application of Intemet technology, lnternet has become a hug information bank ,which includes many kinds of information such as text ,image, video, audio,etc. However ,the topic, format, function and source of image on the Web have variety.it is very difficult to classify images by a unified standard or frame. This paper classifies image as four following preliminary classes based on functionality ： link, decoration, advertisement and content. The preliminary classification can improve the performance and efficiency of image process and can also provide some useful help for other research and application based on Web.

关 键 词：Web图像分类 WEB信息提取 决策树 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]