基于分类的Web图像内容混合自动注释方法  

Automatic Annotation Method Based on Hybrid Web Image Content Classification

作　　者：郑欣[1] 徐军莉[1] 

机构地区：[1]江西科技学院,江西南昌330029

出　　处：《江西科学》2013年第2期232-237,242,共7页Jiangxi Science

摘　　要：Web图像不断增加的数量和多样性,使当前图像检索技术面临很大挑战。然而目前图像检索技术还不能满足人们的需要。陈述一种图像内容混合自动注释方法-HIAM,它包含2个工作模块的Web图像检索,其中一个模块是两个级别的图像分类器将图像转换成图像概念的次序;另外一个是图像语义相似处理模块-ISAM,它具备从网页中挑选与图像概念相关的术语作为图像注释。使传统的文本信息检索技术能够用于图像查询的功能。Increasing amounts and diversities ot web images have made great cnallenge to, current, im- age retrieval techniques. However the present image retrieval technology cannot satisfy people need. In this paper we present a hybrid image annotation model, HIAM, for web image retrieval ,which con- sists of two sub models. One sub model is a two-level image classifier which is used to translate an image inio a sequence of image concepts. The other is image semantic annotation model, ISAM, which has the ability of picking out the terms relevant to image concepts from web pages as image annota- tion. The traditional textual information retrieval techniques are able to use for image queries.

关 键 词：WEB图像检索 图像分类 图像注释 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]