Similarity Measurement and Retrieval of Texture Images
纹理相似性度量研究及基于纹理特征的图像检索

The evaluation of texture similarity is very important in texture synthesis andcontent-based image retrieval systems.In these systems,texture features corresponding tohuman perceptions are more suitable than those features without specific correspondence tohuman perceptions.In this paper,19 texture features corresponding to human perceptionsare used to evaluate the similarity of textures.These 19 features and the correspondingestimation method are proposed based on work by Tamura,Amadasun and Haralick et al.The similarity between different textures is defined by the Euclid distance of their featurevectors.Principle component analysis PCA is used to reduce the number of features from19 to 6 while keeping about 85% capability.Experiments show that the similarity criteriabased on these features works well on large texture image bases.