基于视觉特征的图象搜索引擎设计与实现

Design and Implementation of Image Search Engine Based on Vision Features

文章将基于WWW的搜索引擎技术与基于数据库的图象检索技术有机结合起来,提出了一种基于视觉特征的图象搜索引擎方案,并实现了一个图象搜索引擎系统。该系统先用网络机器人在WWW上搜索图象,提取其颜色、纹理或形状特征,与提交的示例图象特征作比较,把图象的URL按与示例图象的相似度大小顺序存入数据库,然后从数据库中返回结果给用户,并在屏幕上显示。实验结果表明,该系统简单易行,能较好地满足用户的需求。

This paper presents an image search engine scheme based on vision features,which organically combines search engine technology on WWW and image retrieval technology based on database,and realizes an image search en-gine system.Firstly,this system searches images on WWW by Web robot,then,extracts their color,texture or shape fea-tures,and compares them with the submitted image,and stores the images' URLs into database by the order of similari-ty,lastly,returns the results to the users and display on the screen.Experimental results show that this system is simple and satisfies the requirements of the users.