基于重要位平面的彩色图像检索算法

A Color Image Retrieval Algorithm Based on Significant Bit-Planes

针对利用颜色特征检索图像时的不足,提出了一种基于重要位平面的彩色图像检索算法.该算法利用位平面分解的方法对表征图像视觉的信息进行分解,从中提取能够表征图像视觉信息的5个最高位平面,并计算其上的R、G、B颜色分布熵来描述图像特征.为了避免图像中像素颜色值的微小变化对位平面复杂度的影响,采用了位平面的灰度码表示方法.最后设计了以位平面距离加权和的方法进行相似性度量.仿真实验表明所提出的算法具有高效的图像检索性能.

A novel color image retrieval algorithm is proposed based on significant bit-planes to overcome the image retrieval shortcoming from color characters. Firstly, color visional image information is decomposed by bit-plane method, and five highest bit-planes are extracted from the color image by bit-plane theory. Then the RGB color distribution entropies are calculated on the bit-planes, respectively. Meanwhile, the gray-code of bitplane is used to avoid the effect of changes in the image intensity values on the bit-planes. Finally the summation of different bit-plane distances multiplying the corresponding weights is computed to measure the similarity of example and database images. Experiment results show that the proposed image retrieval algorithm is more effective and efficient in retrieving the similar images.