中文摘要：

灰度共生矩阵是一种有效的纹理图像分析方法，为了更有效地进行图像检索，定义了一种新型的灰度共生矩阵描述子，该矩阵描述子能够有效结合颜色、纹理和形状特征，通过描述像素的空间相关性来进行基于内容的图像检索．利用该矩阵描述子进行图像检索时，先在RGB颜色空间中计算彩色梯度，然后利用灰度共生矩阵来描述图像特征．实验结果表明其检索性能优于传统的普通灰度共生矩阵．

英文摘要：

The gray level co-occurrence matrix is an effective method of texture image analysis, in order to improve the performance of image retrieval, the author defines a new kind of co-occurrence matrix descriptors, the co-occurrence matrix descriptors can effectively combine color texture and shape features, the content-based image retrieval is finished by describing the spatial correlation of pixels. Using of the matrix descriptor for image retrieval, the first color gradients is calculated in the RGB color space, and then the gray level co-occurrence matrix describe image features. Experimental results show that the retrieval performance are better than the traditional gray level co-occurrence matrix.