How Facial Recognition Technology Could Help Catch Criminals
One of the growing tech tools law enforcement agencies are using to track down suspects during crime investigations is facial recognition (FR) technology.
In fact, before the FBI reached out to the public Thursday asking for help in identifying the two Boston bombing suspects, they may have tapped FR technology to try and trace the suspects identities, said Jim Albers, senior vice president of MorphoTrust— one of the key suppliers of FR technology to the FBI.
Basically, FR technology is software that uses digital images of a person to match or verify their identity in just a few minutes. An image of a person's face is run through an agency's database to match the image with an identity. U.S. agencies like the Department of Defense, the State Department and the Justice Department all use the technology with their own databases of information for different purposes.
(Read More: Suspects Told Carjack Victim They Were Marathon Bombers )
The technology, though, has big potential to help with criminal investigations — like the developing situation in Boston.
However, it's likely the FBI was unsuccessful in identifying the suspects using FR because either they didn't have a quality image of the wanted persons, or the suspects were not in any of the databases the FBI has access too, Albers said.
"You can speculate that other databases that are confirmed by foreign firms, anything that the DOD had gathered would be available to the FBI," Albers said. "The technology works so well that if they had them in their database, we would've known who the guy was."
While facial recognition technology has high-accuracy when used to match a clear image of a person with another passport-style photo, it is not as effective when used with low-quality images like the ones the FBI released on Thursday. The standard for facial recognition to be accurate requires 90 pixels of resolution between the two eyes of the pictured person. The pictures the FBI released of the suspects were about 12 pixels between the two eyes, said Jim Wayman, the director of the National Biometric Center.
However, Albers said that the technology has come a long way in just the last five years and now there are features that will allow users to clean up even low-resolution photos.
For example, if law enforcement only had the profile of an individual in a picture, they could use MorphoTrust's FR software to create an image of what the individual would look like if his whole face was shown in the picture. The software does this by basically copying the visible side and recreating it on the other side. It won't be an exact match, but it will be close to what the individual actually looks like, Albers said.
"It's all about probability. You are going to get a candidate list and hopefully in the top three you are going to get a match," Albers said.
Law enforcement can then use the created image to run through its database to try and find the person's identity, he said.
While law enforcement may not get an exact match of the pictured suspect — like they would if it was a high-quality, real photo of the individual — they would get several leads for potential matches once they ran the created image through their database, he said.