Centralized Content-Based Web Filtering and Blocking: How Far Can It Go?

To an organisation, centralized Internet filtering and blocking is very important for a couple of reasons. With the flooding of pornographic materials on the Web, educators and parents would like to block these offensive materials from their children. Companies also want to reduce the amount of work time that its employees spend on non-productive Web surfing. Current blocking and filtering mechanisms can roughly be classified into two approaches: URL based and content filtering. In the URL based approach, a requested URL address will be blocked if a match is found in the blocked list. However, keeping the list up-to-date is very difficult. New sites are kept uploading onto the Internet daily; many blocked sites try to use multiple IPS and domain names; the sites might also be moved regularly. In the content filtering approach, keyword matching is often used. Its main problem is the mis-blocking. Many desirable Web sites are blocked because some predefined keywords appear in their Web pages, though in different meaning or context. There are even suggestions for image, audio, and video understanding in real-time content filtering. Of course, the delay time as well as the mis-match between the HTTP streaming protocol and the complexity of the filtering algorithm will be of great concern. In this paper, we investigate how far the multimedia content analysis should go for Internet filtering and blocking. A set of guidelines for defining the heuristics used in the real-time Web content analysis is also given. These heuristics not only have higher filtering accuracy than most multimedia retrieval techniques do, but they also have comparable runtime overhead as the keyword matching does. Our one-year experience of deploying a pornographic filtering system in high schools will also be described. Experience from the system implementation and deployment is found to give a very good direction on the centralized filtering and blocking of Web content.