关键词: 层次匹配, 特征融合, 特征匹配, 灰度共生矩阵, 灰度梯度共生矩阵, 局部二值模式, SURF

Abstract:

Since the efficiency of the lace retrieval method based on image texture features is low, and in order to extract the effective texture features for lace identification, we propose a lace retrieval algorithm containing multiple features fusion through hierarchical matching. Firstly, we process the texture image by the Canny operator and obtain the Canny color image and the gray level-gradient co-occurrence matrix (GGCM). Secondly, energy, average gradient, average grayscale, correlation and other statistical characteristics are used for texture description. Finally, the extracted texture features are matched with geometry features and speeded up robust features (SURF) hierarchically, so the fusion of multiple features under hierarchical matching is realized to compensate for the deficiency of any single matching method and verify the correct rate of the retieval method used in the lace library. Experimental results indicate that the performance of the proposed method is better than other methods, which has a stronger ability of texture identification, and can achieve lace retrieval effectively and improve the reliability and accuracy of image retrieval.

Key words: