融合矢量量化与LBP的图像检索算法

Image Retrieval Algorithm by Vector Quantization and LBP

基于内容的图像检索算法一直是图像领域研究的热门课题,因此提出一种新的融合矢量量化与LBP的图像检索算法。首先,将彩色图像转化到HSI颜色空间,进行矢量量化编码,统计图像码字出现的频数,形成颜色直方图,完成颜色特征的提取;然后,再将彩色图像转化成灰度图像,利用局部二进制模式(LBP)算法提取纹理特征;最后,相似度计算采用颜色特征和纹理特征相似度加权平均,并且改变颜色特征和纹理特征的权值,多次实验得到使查准率最高的权值。实验结果表明,算法能有效地提升图像检索性能。

Content -based image retrieval algorithm is always a hot topic in research of image, and put forward the fusing vector quantization and LBP image retrieval algorithm. Firstly, convert the color image to HSI color space, and carry on the vector quantization, the color histogram is fnrmed by counting up the frequency of each code word to complete tbe extraction of color feature. Secondly, convert the color image to grayscale image, and the image texture feature is extracted by local binary pattern （LBP） algorithm. Thirdly, the calculation similarity use a weighted average of the color feature similarity and texture feature similarity. It is the weight value that is got by changing the weight values of color and texture many times which make precision be the highest. The experimental results show that the algorithm can effectively improve image retrieval performance.