Study Of Query-Based-on-Content Image Database In Web

Image database in Web is a combination of traditional database technique, image processing and Web database technique. The Appearance of it not only broaden the use of traditional database, but also offer a strong tool to support image processing. Web technique makes images database have more applications and many fields, such as medicine, remote sensing, machine, can get images information more easily. Then many fields can share images resource more expediently in Internet.This paper mainly aims at images database system based on Web. The new image database system in Web should query by image content, it relates to image processing, distribute object technology and component object model technology, Web database, connectivity database technology and database middleware technology and so on. The paper lucubrates some characteristics and framework of the distributed multi-tier applications; and discusses many kinds of characters, structures of component object model such as COM, DCOM, COM+ etc.; and expounds web database technology in image database system relating to choose database, connect database (such as ODBC, JDBC ) , database middleware ( such as ISAPI, ASP, JDBC , CGI, etc. ). The paper describes how to select ODBC interface and CGI middleware based on the requirement of system and the practical environment. The paper studies and demonstrates hybrid structure of image database system Querying By Image Content (QBIC) in B/S mode and its framework relying on the basic theory of the distributed multi-tier applications, component object model and Web database technology.How to query by texture of image in Web have also been studied in this paper. There are two methods that have been mentioned: (1) using distance of spatial histogram to measure the similarity between images; (2) using distance of a composed of four features detected from spatial dependent matrix to measure the similarity between images;At last, how to query by color of image in Web is discussed. There are also two methods that have been mentioned: (1) using distance of RGB histogram to measure the similarity between images; (2) using distance of HSV histogram to measure the similarity between images.The system of image database querying by image content in Web has been developed, which are composed of the image preprocessing, querying by the texture of image, the color of image methods mentioned above and querying by text.