Not just for law enforcement anymore: The evolution of biometrics
04 May, 2015
category: Biometrics, Digital ID, Government, Library
When it comes to solving a crime law enforcement officials ask many question. Was there an eyewitness? Was evidence left at the scene? What about latent fingerprint images?
Other than names, fingerprints may be the oldest identifier. The Babylonians realized fingerprints were unique and used them for identification more than 3,000 years ago. But only in the last 100 years did the biometric truly become a science used to identify criminals and conduct background checks on individuals working in schools, social work and other potentially sensitive positions.
In the past 40 years, fingerprint science has moved from examiners studying ink-stained sheets with magnifying glasses to Automated Fingerprint Identification Systems (AFIS), which use computer systems and software to pull up possible matches. This does not remove the human aspect as examiners still compare the actual fingerprint image to the AFIS-selected matches.
AFIS standards still emerging
Automated Fingerprint Identification Systems (AFIS) have been around for 40 years, but standards for these systems didn’t emerge until 1993, says Robert Horton, senior director of marketing and communications at MorphoTrak.
The first efforts to standardize AFIS started in 1986, but as with most standards it took time to work though the process. The first standards were “loosely written” so that older system were grandfathered in and didn’t’ have to make changes, Horton explains.
Since 1993 the standards have been modified every five years by the National Institute of Standards and Technology. The image-based standard calls for systems to extract minutia points out of an image in order to make a match.
But even the standard isn’t perfect and it can be open to interpretation, Horton says. The specification primarily searched fingerprint ridge endings, but it doesn’t specify where a ridge ending begins or ends. This can cause complications when sending prints to the FBI, which had to create translation software so the system can see where ridges begin and end in order to read the fingerprint the same way.
Work is underway to solve the issues with these standards and make it easier to transmit and search different AFIS, a step that could lead to greater interoperability between jurisdictions.
If the procedural crime drama is to be believed, however, these prints are scanned and run against a single national database with a match returned in seconds. But as with most things on television, the reality is quite different. The majority of AFIS are siloed, such that bordering states and jurisdictions typically can’t even share data.
While connectivity has always been an issue with these systems, the technology is evolving.
Though AFIS has fingerprint in the acronym, it’s beginning to incorporate much more than just the single biometric modality. Systems are starting to collect facial images, palm prints, irises and even tattoos to assist in identifying criminals. The term AFIS is being used as a catchall for the use of biometrics to identify criminals.
What is AFIS?
AFIS technology uses a combination of fingerprint images and templates to identify an individual from a database of millions, says Robert Horton, senior director of marketing and communications at MorphoTrak. Fingerprints images are scanned and the minutia points – ridge endings and bifurcations – are translated into templates so they are easier to search in a database. In essence, the template is a numeric representation of key geographic points from a unique fingerprint image.
Both the actual fingerprint image and the resulting template are retained in separate databases, Horton says. The AFIS searches the templates to find likely matches, but the final comparisons are done with the actual images.
When a local jurisdiction sends information to the FBI or another agency, it’s the images that are transmitted. Jurisdictions send images back and forth because different AFIS use proprietary templates and algorithms that can’t be used by competitors. By sharing the actual images, each AFIS can create its own template to use in its specific search.
These systems are not intended to return conclusive matches but rather a list of likely candidates. At this point a human examiner manually compares the actual fingerprint images identified in the AFIS candidate list. A newer AFIS can make this a much quicker and simpler process by returning more accurate results, Horton says.
These newer systems cut down on the time it takes to find matches and can help law enforcement agencies to better allocate resources, Horton explains. Examiners spend less time manually checking prints because it’s more likely that the match will be found in the top few candidates provided by the system.
The Feds go next gen
On the cutting edge of criminal identification will be the FBI’s Next Generation Identification (NGI) system. The system is about three-quarters of the way through its implementation, says Stephen Morris, assistant director of the FBI’s Criminal Justice Information Services Division (CJIS) in Clarksburg, West Virginia. The NGI will replace the Integrated Automated Fingerprint Identification System (IAFIS) that has been in use for about 20 years.
“It’s been 10 plus years since we started working on the NGI, recognizing that we would need to upgrade the technology, leverage modern technology, as well as provide some enhancements and some additional service to law enforcement agencies,” says Morris. “The legacy IAFIS system was the FBI’s first step in providing some sort of automated capability for law enforcement agencies to submit fingerprint data, do some comparisons and get responses back. NGI represents the next generation of that.”
CJIS maintains the FBI’s criminal history and fingerprint databases. It shares information with 18,000 law enforcement agencies in the U.S., as well as national and international security agencies.
Morris says the NGI program is a major upgrade. “We’ve been able to provide better search algorithms that not only improve accuracy, but also give us a lot more capabilities particularly as it relates to comparing latent fingerprints,” Morris says. “Essentially, it represents better, faster and smarter technology.”
The FBI’s Next Generation Identification system is designed with change in mind. This includes the addition of new modalities such as DNA, iris and facial technologies
The new system is designed with change in mind. One of the big drivers behind the program is recognizing that 20 years from now, the technology will be different. “There’s going to be new modalities of biometrics such as DNA, iris, facial – technologies that aren’t used everyday or are just coming into use,” Morris says. “We wanted to make sure that the system was scalable and that we could upgrade it and use a plug and play capability.”
