摘要
为了解决现有颜色直方图检索技术所存在的鲁棒性较差、特征维数过高等问题,本文提出了一种基于位平面颜色直方图的图像检索新方法.该方法首先结合噪声攻击特点(对高位影响较小),提取出重要的位平面图像并确定其视觉权值;然后构造出位平面图像颜色直方图;最后以位平面图像颜色直方图为特征,计算图像间内容的相似度并进行检索.仿真实验表明,本文方法不仅能够准确检索出用户所需图像,而且对光照、锐化、模糊等噪声攻击均具有较好的鲁棒性.
In order to overcome the deficiency of existing color histogram based image retrieval,a robust color histogram image retrieval based on bit-plane is proposed in this paper. Firstly,the important bit-plane (high bit-plane) images are extracted from the origin color image according to the characteristic of noise attack,and the relevant perceptual weight values are assigned. Then .the color histograms of bit-plane images are built. Finally,similarity between images is computed according to the color histogram of bit-plane images. Experimental results show that our image retrieval is more accurate and efficient in retrieving the user-interested images. Especially,it can retrieval the noise image effectively.
出处
《小型微型计算机系统》
CSCD
北大核心
2007年第4期715-719,共5页
Journal of Chinese Computer Systems
基金
辽宁省自然科学基金项目(20032100)资助
视觉与听觉信息处理国家重点实验室(北京大学)开放基金项目(0503)资助
大连市科技基金(2006J23JH020)资助
"图像处理与图像通信"江苏省重点实验室(南京邮电大学)开放基金项目(ZK205014)资助
江苏省计算机信息处理技术重点实验室开放课题基金项目(苏州大学KJS0602)资助.