Research On Image Resource Retrieval Technology Of Digital Library Based On Non Sampling Contourlet Transform

Image can describe objects objectively and accurately.It contains rich visual information,and it is an important information carrier in human social activities.People can understand and recognize things more visually and visually through images.With the development of computer technology and the popularity of the Internet,there are more and more kinds of digital images,and the number of digital images is increasing with each passing day.Faced with this situation,how to make users find the required images accurately and efficiently in large-scale databases has become an increasingly serious problem.Therefore,image retrieval technology arises at the historic moment.At present,there are mainly two ways of image retrieval research,one is text based image retrieval(TBIR,Text-based Image Retrieval),the other is content based image retrieval(CBIR,Content-Based Image Retrieval).Based on the study of image retrieval of the background,research status and key technology content,using bibliometric methods to analyze international image retrieval literature published in the past fifteen years,explores the development of image retrieval in the field of literature from the time distribution,subject distribution and other aspects,and focus on the author,institution,state the key point of international mining and word retrieval major research efforts and research hotspots,and provide a theoretical basis for further promoting the domestic research.In the era of big data,users' information literacy is getting higher and higher,and the quality requirement of information is also improving.As the main provider of information,library is responsible for the accurate and fast transmission of high-quality information service.With the continuous maturation of image retrieval technology,the application and development of digital library image retrieval technology has come to a turning point.In order to improve the quality of library information service and meet the needs of information age,based on in-depth study of the theory and development of image retrieval,this paper constructs an efficient image resource retrieval model framework for digital library.The model is composed of five parts: information acquisition module,image preprocessing module,feature extraction module,relevance feedback module and similarity measurement module.In the specific implementation process,the author adopts the method based on non sampling Contourlet transform(NSCT)texture feature extraction algorithm,this method compared with other extraction methods of texture feature can obtain more multi band coefficient information,while the cancellation of the down sampling operation,the implementation of digital library image retrieval.The experimental results show that the method has good retrieval performance,high precision and recall.The digital library image retrieval framework proposed in this paper is feasible,and it can provide users with high quality search service,providing a reference value for the development of Digital Library's retrieval technology.