基于颜色共生矩阵的纹理检索算法MCM

描述了一种基于颜色共生矩阵的纹理检索算法MCM,主要包括颜色共生矩阵纹理特征提取算法以及纹理特征的相似性度量函数,给出了利用MCM算法检索图像库的实例.通过MCM算法提取的特征除了反映图像的纹理关系外,还综合了其颜色构成特征,部分建立了与人的视觉感知之间的对应关系.实验表明,MCM算法优于一般的灰度共生矩阵纹理检索算法,并且具有较好的检索效果.

An Image Texture Retrieval Algorithm Based on Color Co-occurrence Matrix (MCM)

Content-based image retrieval (CBIR) is one of the most important researches currently focusing on the combination of multimedia, information retrieval, artificial intelligence and database. As the common feature of image, texture analysis is a key research of CBIR. In the paper an image texture retrieval algorithm named MCM is presented, which includes algorithm for extracting texture features from color co-occurrence matrix and similarity measurements. On the basis of gray level co-occurrence matrix and image retrieval algorithm in use of dominant colors of image sub-blocks, the concept of color co-occurrence matrix is described. Then, computed on color-connected regions, the color co-occurrence matrix represents not only image texture characteristics but also color structure. We give some examples using MCM retrieval algorithm. In these examples, the results are images with similar texture coarseness and distribution. Moreover, the ending images have the color similarity to the original ones. We also give some experimental results using different retrieval algorithms, which indicate that MCM achieves good effect and takes evident advantages over the traditional gray-level-co-occurrence-matrix-based algorithms, especially for images whose color and texture are both crucial.