Digital Library-Retrieval Of The Document Imaging

Along with the rapid development of digtal library,the method of the Library information collection,processing,MARC records retrieval etc.has been hard to meetthe needs of users.Information retrieval technology has been developed from field retrieval to the content retrieval,and search engine.Most of them are documents image.This article is based on the content of the document images in the retrieval of the study.It can automatically extracted image contents as searching characteristics through the machine.And then use similarity algorithm approximate matching.This article mainly discourses from two aspects.Firstly,discuss the principle and the modeling of the documents image.Combining the principle and the retrieval modeling,research the each major part in retrieval system.Then,explain the retrieval platform from every point of view.This system has four main retrieval algorithm of composition:Document image preprocessing,Chinese image segmentation,the feature extraction and matching,and Client retrieval.Pretreatment part,this paper adopts threshold value method of document binary image.Segmentation,Based on the characteristics of the image using projection method,Segment Chinese character image into single image block.Do the thinning to the character using the improved template.Burr was significantly reduced.This paper mainly introduces six kinds of feature extraction using a carcass traits and characteristics of matrix feature extraction and secondary matching strategy.That recognition and recognition rate is relatively ideal.Client retrieval,this system Can achieve word class retrieval.Output the retrieval results according to the similarity of the output.Using VC++platform,do the retrieval in the image database,it Can be seen that the search rate is higher.