中文摘要：

提出了一种基于分块颜色体积核特征和改进的局部二元模式（MLBP）的图像检索算法.首先对图像进行重叠九分块并提取子块的颜色体积直方图,并使用高斯核函数将低维颜色特征映射到高维核空间对图像先进行过滤,再使用MLBP算子对返回的图像提取纹理特征,最后对特征进行相似性度量.实验结果表明,相对于以颜色体积直方图、分块颜色直方图、传统LBP算子为特征的图像检索方法,该算法有效地提高了检索结果的精度、改善了相关图像的排序值,并具有很好的抗噪性.

英文摘要：

In this paper, a method of image retrieval based on block color volume kernel feature and modified Local Binary Pattern （MLBP） is proposed. First the image is divided into blocks and nine overlapping sub block histogram to extract color volume, and low dimensional color feature mapped into high dimensional kernel space of image before filtering using the Gauss kernel function, and then extract texture features of the returned image using MLBP operator, finally the feature similarity measure. The experimental results show that, compared with the image with color volume histogram, block color histogram and LBP operator for feature retrieval method, The algorithm can improve the accuracy of retrieval results, improve the sorting value of related images, and has a good anti noise performance.