"The system functions as a search engine," said Quan Zhang, chief program scientist. "You input a website address, and it will automatically open every page on the site and examine the words, sentences and paragraphs for pornographic materials. It is smarter than keyword-based search engines because it can discern articles which are critiquing porn materials from those which are merely talking dirty."
The system, developed by the CAS Institute of Acoustics last December, recognizes Chinese characters and identifies porn through keywords, syntax and language index.
The only problem with the system, according to Zhang, is that the system can only recognize "pornography in written form," and that, "We have yet to find an effective way for audio-visual examination, which needs further advanced technology."
Even in its present form, the system is expected to be helpful in the government's ongoing crackdown on Internet pornography, which has led to 1,911 website closures and 41 arrests since its start on Jan. 5.
"We anticipate that many government organizations and website operators will find it useful," Zhang said, "and we will soon put the software into the market."
The system is already in trail use on the websites of CAS and some of its institutes.
