基于纹理特征的图像检索方法的研究与实现

在研究与分析了纹理特征提取算法的基础上,设计并实现了一个图像检索实验系统。该系统相似性度量采用欧氏距离方法,特征提取算法为共生矩阵,并且利用UIUC图像库中五类536幅蝴蝶图像和Caltech101图像库中的barrel,dalmatian和sunflower三类等图像数据库验证了共生矩阵提取纹理特征的检索方法,并对不同测试图库所得到的实验结果进行比较。从实验结果可以看出,对于具有明显的纹理特征的图像数据库利用共生矩阵提取算法还是具有较好的检索性能,可以得到较好的检索效果。

Research and Implementation of Image Retrieval Method based on the Texture Feature

The paper has research and analyzed the arithmetic of texture features extraction,the system of image retrieval are design and implemented.The system adopted the Euclid distance metric,and the last,according to the image database of the peacock which from UIUC and the database of barrel sunflower from Caltech 101,the Co-occurrence matrix features extraction are validated. Compared these experiment results,at the different database,using the arithmetic of Co-occurrence matrix features extraction,some image which have the obvious texture feature can be obtained the better retrieval effect.