Facial-recognition technology is used primarily to confirm a person's identity or conduct background checks by matching that person's facial features with facial-contour data stored in a database. The applications of facial-recognition software are significant.
For the government, this means that different areas of the Homeland Security Department can mine law-enforcement databases in order to flag illegal immigrants or suspected terrorists attempting to enter the United States.
In the transportation industry, airlines are starting to deploy facial-recognition technology to meet the International Civil Aviation Organization's plans to integrate biometric identification information into passports and other machine-readable travel documents. The organization in mid-2003 chose facial recognition as the de facto standard for machine-assisted identity confirmation, saying in a press release at the time, "the face rated highest in terms of compatibility with key operational considerations, followed by fingers and eyes." Chopra also notes that more than 150 casinos use facial-recognition technology with surveillance equipment as part of their security practices.
Although facial recognition isn't widely used across a great many industries, the prevalence of 2-D systems among those employing facial recognition could be an impediment to the introduction of 3-D. "3-D face-recognition systems require a comparison of 3-D images with 3D images," Chopra says in her E-mail. "In order to capture 3-D images, special sophisticated cameras are required, which are much more expensive than 2-D cameras."
Yet deficiencies of 2-D facial recognition are likely to push 3-D adoption. 2-D facial-recognition technology is less accurate in uncontrolled environments where the position of the face and the lighting aren't optimal, Chopra says. 3-D is more accurate because it corrects for different poses and lighting variances.
One solution is to try "2.5-D" methods, so named because they compare 2-D images in current databases with synthesized 3-D images. The hardware used is the same as 2-D, but software upgrades are required to allow for the 3-D modeling, Chopra says. Companies and government agencies have to weigh 3-D's improved accuracy against its higher costs. 2-D won't become obsolete until the cost issues are resolved.
Still, just like Jaws, 3-D is coming. A4Vision Inc. this week revealed an investment and development deal with CIA-sponsored In-Q-Tel to advance A4Vision's 3-D facial-scanning and -recognition software and equipment. Even traditional 2-D facial-recognition companies such as Viisage Technology and Identix have either developed or are in the final stages of their 3-D facial-recognition technologies, Chopra adds. Other 3-D facial-recognition companies include 3dbiometrics, Geometrix, Neurodynamics, and Genex Technologies.
All of this activity is expected to help create more than a $4 billion market by 2009 for fingerprint, facial-recognition, iris-recognition, hand-geometry, voice-verification, and signature-verification biometric technologies, according to Frost & Sullivan's World Biometrics Market study. International Biometric Group, a biometric security consulting and technology services firm, is even more bullish, predicting biometric hardware and software sales will reach $4.6 billion in 2008, up from an estimated $1.9 billion in 2005.