一种基于图像内容的图像评价方法

Content-based Image Quality Assessment

根据图像内容质量对图像进行区分处理,可以有效地提高图像搜索效率。本文提出一种图像评价方法,专注于改善不同质量图像对搜索效率的影响。该方法基于图像颜色丰富程度、局部特征丰富程度以及是否背景丰富、是否包含多个子图等信息进行综合质量评价处理。实验结果表明,该方法时间效率很高,可以和图像搜索系统有效结合,有效改善图像搜索系统的搜索召回率和准确率。

Content-based image retrieval distinguishing in accordance with the quality of the image content is one of present image retrieval development,and it can effectively improve the efficiency of image retrieval.In this paper,one method of image quality assessment is proposed,which is focused on improving the image of different quality impact on the retrieval efficiency.And,this method,based on the color of image,local property,background and sub-image,takes a comprehensive quality assessment processing.The experimental results show the following three conclusions：（1） high time efficiency is achieved by this method;（2） this method can combines with image retrieval system effectively;（3） searching recall rate and accuracy rate are significantly increased by use of this method.