Canadian cops explore biometrics to swap mug shots

An Ottawa-based company has developed a networked search technology designed to cut the time it takes for police to identify crime suspects from 24 hours to less than 10 seconds in some cases.

In the first pilot project of its kind in Canada, according to BlueBear Network International, three Ontario law enforcement agencies have spent the last two years testing out the biometrics software maker’s Integrated Digital Law Enforcement (IDLE) system. Now available to law enforcement and justice agencies worldwide, IDLE allows members of the police community to share so-called electronic mug shot databases over the Internet.

The project, a partnership between BlueBear and Sun Microsystems of Canada Inc., linked police officers in Chatham-Kent, Windsor and York Region via a virtual private network that delivered face-recognition search engine software from VisionSphere Technology Inc. Powered by Sun Fire V20z servers, IDLE runs on the Linux operating system at the stations and the Sun Sparc Solaris platform at a central station that houses the network guardian. The location of the station has not been disclosed to the public, according to a Sun spokesperson.

After discovering law enforcement agencies’ need for information sharing and their under-utilization of electronic mug shots, BlueBear chief executive officer Andrew Brewin worked with his company to develop biometrics software for police. “Ultimately it ends up being a joint effort,” said Brewin. “When someone is arrested and booked, their face will be searched so people can validate their identity.” Brewin added the technology could also be applied to investigation and detective work in terms of identifying victims through a driver’s license database, for example.

IDLE takes the physical image like a photograph, video clip or fingerprint and creates a code or a template that is read by an algorithm. The code is then used in the search to generate a score based on how likely the picture matches that in the database.

Detective sergeant of York Region’s forensic identification bureau David Juck said IDLE uses existing police databases, which store suspect photos and fingerprints, more effectively than before. In one case, for example, York Region did 148 searches and found 73 matches, said Juck, adding the system was accurate about 50 per cent of the time in York and 67 per cent of the time in Chatham-Kent.

“We liked the speed and accuracy of the system and how it makes better use of your database,” said Juck. “We could never do this before.”

The system also allows police to use surveillance video footage and composite sketches to identify suspects.

Because all of the separate databases are connected, agencies don’t need to combine all the information into one database, said Brewin. While the technology enables police organizations to share information with each other, it also presents organizational issues, Brewin added. “There’s a huge demand for it,” said Brewin, citing prospective customers in a couple U.S. states as well as Turkey, Singapore, China and India. “The ironic part to all of this is, at varying levels, the police all want to share information with each other. Something seems to happen that causes them not to.”

In addition to using biometrics to identify suspects, the system also uses face recognition technology for officer log on. Once an officer is logged on to the system, for example, another officer can’t start searching for information, said Brewin.

Likewise, Norman LeCouvie, executive director of global government partners and solutions at Sun Microsystems, said there are sovereignty issues around the sharing of data. “We wanted to make sure that the system didn’t interfere with those organizations,” said LeCouvie. “The system is now such that one city can peruse on its own database but ask that the facial template exists in the other two cities.”

Juck was initially skeptical about trying the software, which will be used in conjunction with more traditional identification methods like line-ups and age and descriptor searches. “My original reservations were that it was just hocus-pocus,” said Juck. “Until you do the tests yourself, I just didn’t believe it.” Juck, however, is slightly hesitant about using drivers licenses to identify victims due to the sensitivity and legality issues.

Chatham-Kent and York Region are considering implementing the system at their sites in the upcoming months. Windsor is also evaluating the technology for future use.

Comment: info@itbusiness.ca

Share on LinkedIn Share with Google+