Feature extraction is the basis of image retrieval,and the choice of its algorithm is directly related to the retrieval performance of the system.The article explains the two algorithms of color histogram and Hu invariant moments in detail,and since image retrieval with single features does not meet people's needs,the article conducts a deeper research on image retrieval based on feature fusion.In order to make a better fusion of color features and Hu features,Hierarchical Analysis is added to set the weights of single features.The experiments show that the method of introducing hierarchical analysis to set the weights of two single features not only improves the retrieval performance of the image retrieval system,but also improves the retrieval speed and retrieval efficiency.
万方数据
