基于重要位平面的真彩色图像检索算法  

Color Image Retrieval Algorithm Based on Significant Bit-Planes

作　　者：田宏[1] 杨树刚[1] 

机构地区：[1]大连交通大学软件学院,大连116028

出　　处：《计算机辅助设计与图形学学报》2010年第2期279-285,292,共8页Journal of Computer-Aided Design & Computer Graphics

基　　金：辽宁省自然科学基金(20072157)

摘　　要：在研究利用图像颜色特征和颜色的空间分布特征进行图像检索的基础上,提出一种基于位平面理论的真彩色图像检索算法.为了更好地描述图像的检索特征、优化检索性能,定义了位平面一阶矩、位平面平坦度和位平面的凹凸面3个概念,并依据这些概念提取出了图像的颜色特征和颜色空间分布特征;同时设计了与所有提取出的图像检索特征相适用的相似性度量准则.仿真实验结果表明,该算法具有较好的检索性能.A into account describe the introduced, color image retrieval algorithm based on the theory of bit-plane is proposed, which takes the color features and the properties of spatial distribution of the color. In order to better retrieval features and to improve the that is, bit-plane mean, bit-plane flatness retrieval efficiency, three new concepts are and bit-plane roughness. Based on these three concepts, similarity proposed the features and measure scheme is spatial distribution of the designed for these retrieval image retrieval algorithm is more effective and color are extra features. Experi efficient in retrie cted. Finally, a reasonable mental results show that the ving similar images.

关 键 词：基于内容的图像检索 位平面理论 位平面粗糙度 灰度码 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]