Content-based Stamp Image Retrieval

The stamps,which originated in France in the mid-17th century,were first used as a postage certificate.In today's society,with the rapid development of electronic communication technology,the role of stamps itself has been gradually diluted,and the value of its collection is more important.At present,there are very few databases on stamps,and for the retrieval of stamp images,it is text-based image retrieval,namely,retrieval stamp image by the name or the description of the stamp.Therefore,the research and development of content-based stamp image retrieval system is very necessary,which is the main purpose of this paper.The content-based image retrieval is mainly divided into two parts,one is the extraction of image features,and the other is the feature matching of images and the calculation of image similarity.In this paper,according to the characteristics of the stamp image,targeted introduced the extraction of the extraction of color features and texture features,this paper proposes a combined with color features and texture features of image feature fusion image retrieval method.The experiments of stamp image retrieval are carried out by using three methods,and the corresponding recall and precision are calculated as the criteria for evaluating the retrieval methods.The experimental results show the recall and the precision of the multi-feature fusion method are higher than that of retrieval with single feature.In order to make it more convenient for users to use stamp image retrieval system,this paper uses MATLAB software to build a user interface under Windows 7 system.The user can upload the stamp image to be retrieved and view the retrieval result.The related feedback mechanism is also added to the stamp retrieval system.When the stamp image is not related to the image to be retrieved,the system will adjust the weight of the retrieval feature and retrieve it again to get the new retrieval result.