一个基于对象的图像检索系统:Basestar

图像检索是近年为适应国际互连网以及图像数据库高速发展而出现的一门新技术。该文建立了一个基于对象的图像检索系统-Basestar。该系统针对传统图像检索方法偏重图像整体特征,忽略用户对图像感知等缺点,采用用户参与方式,自动准确分割对象,并在此基础上利用对象颜色和形状特征对图像实现检索。实验结果表明该系统取得了良好的检索效果。

A New Image Retrieval System Based on Object:Basestar

By the growing of image database and Internet,multimedia data grow startling.How to manage and index these data becomes a current urgent question.Image retrieval technology is introduced in this area.In this paper a new image retrieval system based on object named as Basestar is discussed.Many image retrieval systems pay more attention to the whole image features.However Basestar system pays more attention to human perception.Basestar can segment the object based on the user's perception.Based on the segmentation result,the image features-color and shape of the segmentation result are extracted to compose of image feature vector.By the experiment Basestar system has better retrieval performance.