Video Surveillance Blog

« IP video to double by 2011 | Main | Cameras Look to Catch Graffiti Artists in the Act »

March 23, 2007

Facial-Recognition Helping Banks Curtail Phony Checks

The effectiveness of facial-recognition technology in video surveillance has been cause for debate in the past. Some claim the technology -- originally seen as an advanced tool in the fight against terrorism -- is overhyped and has yet to produce solid results. That's not the case in banks however, where video facial-recognition systems are proving to be useful in combating check fraud.

64,000 cases of check fraud were reported to the Treasury Department last year. It's a serious problem, and the type of people who pass bad checks are often repeat offenders. This is one of the reasons why many feel video surveillance systems equipped with facial-recognition technology can cut down on such activity. One company pushing this technology is 3VR Security Inc., a San Francisco based video security firm.

Here's how the 3VR facial-recognition system works: If someone reports that a bad check has been cashed on their account, the 3VR system collects information about the time of the transaction, and the teller station, then uses that information to search the bank's digital video recorder for images of the check casher at the teller window. Additional footage from the bank's other security cameras is also used to get the best overall image of the offender. This image is then stored on file. The next time the culprit arrives at the teller window, his or her likeness is captured on camera and compared to the saved image. The system then sends the original image to the teller's video screen. If the likeness matches the customer, the teller can alert security.

Another feature of the 3VR system is that it can compare the fraudster's picture to all other check cashers who've come through the bank. If the suspect is spotted on video at a different time, the account information involved in that transaction is noted, and the account holder can be contacted to determine whether the check was authorized.

So far, banks using this system are hailing it as an extremely valuable tool. One of the reasons facial-recognition has been somewhat disappointing in other applications is its inability to effectively match database images with real-life pictures of people in motion within crowds. But at banks, where lighting conditions are consistent, and customers are standing still at the teller window, it's much easier to accurately match up likenesses and identify likely suspects. Based on the positive response from banks currently using the system, it appears that facial-recognition video surveillance could become a major security tool for many more banks in the near future.

Bookmark this in Del.icio.us.

Posted by Dan on March 23, 2007 10:35 AM

Permalink | Comments (0)