图像表示及在检索系统中的应用

Graph presentation and its application in information retrieval system

目前,网上流行的过滤和搜索系统主要是依靠文本内容,很少有图像搜索系统。广义的图像搜索引擎可以输入图像,从网络中搜索类似的图像及文本内容,免去了语言的障碍。常用的技术为基于内容的图像检索(CBIR),它是图像分析的一个的重要研究领域,也是目前搜索引擎技术的难点。造成图像分析与检索困难的原因是目前还没有标准的描述图像的方法,众所周知,图像相对于文本理解起来更为复杂,表达的信息也更为丰富。本文阐述了一种基于标题、颜色、形状等多种特征的图像信息表示方法,并利用文本分析、颜色直方图的颜色特征提取、图像边缘检测实现了对图片信息的展示与检索功能。

Nowadays,the popular Internet filtering and search system relies mainly on text content,there is little image search system.Broad image search engine＇s input is image,and its output are similar images and texts,eliminating the language barrier.Commonly used technology--the content-based image retrieval（CBIR）,which is an important research field of image analysis,is also difficulties in the current search engine technology.Difficulties caused by image analysis and retrieval are that there are not standard methods of image description.As we know,the image description is more complex,the information of image is also more abundant.In this paper,we developed a new method based on the title,color,shape and other characteristics of the image to represent am image,and use the methods of text analysis,the color histogram feature extraction,edge detection algorithm to achieve the display of images and retrieval of information.