A pilot project launched in three southern Ontario police forces is using distributed networked search technology and a facial recognition search engine to bring down the bad guys.
The project, currently being tested within the Chatham-Kent, Windsor and York Region police services, allows participants to share and search each other’s mug shot and text databases simultaneously and securely over the Internet. This is done using VisionSphere‘s VS-Ident solution, which its CEO, Ottawa-based Sal Khan said has the potential to connect every police department in Canada.
“”There are approximately 500 police departments in Canada and if they were all linked together we would have a total of four to five million mug shots that could be searched simultaneously,”” Khan said.
If adopted on a large scale after the pilot project is completed, larger police departments with arresting and booking systems can expect to benefit from the shared information, while smaller departments could potentially have access to a mug-shot database for the first time via the Internet, Khan said.
Dave Juck, York Region’s point person on the pilot project, said that he is optimistic about the use of the technology.
“”If someone has been assaulted and works with a composite artist to come up with a composite sketch, it can be put into the machine and we’d expect to see some kind of results,”” he said.
The technology compares faces using a template, which Juck said distinguishes hard features such as eye color or nose width rather than race or sex.
“”It takes the topography of the face and finds possibilities to look through,”” Juck said.
VS-Ident can also search databases using other image sources including surveillance videos and security photos. According to Khan the system takes mere seconds to sort through the images, which is an important factor.
“”Think of what’s in the news right now,”” he said, referring to Toronto’s current investigation of the murderer of 10 year old Holly Jones. “”If the police had the description of a suspect, they could run it through thousands of images in seconds.””
The pilot project is using Sun Microsystems‘ Sun LX 50 servers running Linux, and in several preliminary tests it was reported that one of the servers searched 100,000 facial templates in four seconds over the Internet.
According to Khan, the tool could be used in other environments where photo databases are kept such as border crossings . While it has the potential to administer databases of other catalogued information like fingerprints, Khan said there are no plans to include such information at this time. However, the search engine is able to search text databases that house information such as tattoos and distinguishing scars.
Juck sees the technology as having great potential for the future of law enforcement, but stressed that it is just another tool.
“”If it’s going to be effective, we’ll add it to our arsenal of tools,”” Juck said. “”Anything to help bring a criminal to justice —— that’s the name of the game.””
The project is sponsored by the Canadian Police Research Centre.