Facial recognition software set to boom
Face recognition technology could provide an answer to growing security concerns, according to Dubai-based Business Automation & Security Systems (BASS).
Face recognition technology could provide an answer to growing security concerns, according to Dubai-based Business Automation & Security Systems (BASS). Following September 11th, many companies have been rethinking corporate security, and there is growing interest from governments in biometrics. Now, with improvements to face-recognition software, face ID looks set to become the predominant biometric solution said Jamie Brooker, biometrics manager at BASS.
“Fingerprints have been associated with criminal detection for over one hundred years, so people are reluctant to use fingerprint recognition; there are similar connotations with iris scanners, and people are nervous about their eyes,” said Brooker. “Also, you have to learn how to put your fingerprint on to a scanner—it sounds silly but you have to address the lowest common denominator [with biometrics]. Fingerprint recognition is 10-20% more accurate when the user has been trained—but you don’t have to be trained to look into a camera.”
BASS is promoting Visionics FaceIT facial recognition software, which is primarily intended for deployment in two roles—surveillance and authentication. The technology works by mapping over 300 different differentials on a two-dimensional, greyscale image of a human face, such as the spacing of the eye sockets, and then comparing an algorithmic template of that face against a database of recorded faces.
The system can scan and match up to 150 faces per second running on 600MHz, 256Mb RAM PC, with an error rate of less than 1%. Because of the way the system samples faces, it is able to cope with things such as facial hair, glasses and other changes to appearances that may change over time, and gives equal performance on all races and genders.
For authentication purposes such a door controls or computer logons, the system provides a much cheaper and convenient form of control than passwords or swipe cards, or other biometric solutions, said Brooker. Users only have to face a camera to activate the recognition technology, rather than use retina scanners that some users may be squeamish about, or fingerprint scanners that take some training for users to learn how to use properly.
“A swipe card or password anyone can use, but only one person can use a face, and it is a lot more convenient from the users point of view, and the administrators point of view, as you don’t have to worry about lost cards or forgotten passwords,” he explained.
The system is also a lot cheaper to administer that a standard card, key or password based authentication system, as once a face has been sampled, it does not need to be changed. The technology involved, such as a simple web cam, can also be used for functions like video conferencing alongside authentication purposes.
For surveillance purposes, FaceIT can be used to unobtrusively scan faces in environments such as shopping malls or airports, which can then be compared against a database to screen for undesirables or missing persons. FaceIT can identify faces from drawings, artist impressions, e-fits or video capture, and although the set up is more expensive, and processor intensive than authentication, it is gaining acceptance. “The IT and security people realise there is a need for greater security, and we are marketing the solutions to all segments—it is just education more than anything else to convince the people at the top,” Brooker said.