Tattoo Recognition: Law Enforcement’s Newest Identification Tool
by David Reutter
New technology is giving law enforcement agencies the ability to identify people by taking a photo of their tattoos; it can also group people with others who have the same type of body art.
Federal researchers at the National Institute of Standards and Technology (NIST) have launched a program to accelerate tattoo recognition technology. In 2014 and 2015, the Institute initiated its Tatt-C program, which stands for Tattoo Recognition Technology Challenge.
The project started with an FBI database containing 15,000 images of prisoner tattoos. The biometric database was shared by NIST with 19 participating organizations: five research institutions, six universities and eight private companies; the objective was to use their algorithms to create a database for law enforcement officials.
Five tests were involved in the Tatt-C project. The first was determining if an image contained a tattoo. With a reported 90% accuracy rate, it is evident that technology is already highly sophisticated. The highest result came from biometric technology company MorphoTrak.
Next, the participants were asked to match images of the same person’s tattoo. Overall, they reported more than 95% accuracy. MorphoTrak reported an accuracy rate of 94.6%, while Purdue University reported 91.6%.
Another test required matching tattoos to a similar image in another medium; in other words, matching tattoos to street graffiti or a drawing. Here, the results were poor with the best reported accuracy coming from the MITRE Corporation, at 36.5%.
There was also a tattoo similarity test, which was the most controversial for its ability to group people according to their religion, politics or other affiliations based on their tattoos. In this test, the participants grouped people with tattoos of religious symbols such as crucifixes, praying hands holding a rosary or gang symbols. The results were very poor, with MITRE reporting the highest accuracy rate of 14.9%.
Tattoos have long been recorded by law enforcement officials during arrests or incarceration; for example, mug shots of arrestees’ body art have been taken since the turn of the century. Typically, photographing tattoos resulted in investigators having to spend hours going through telephone book-size portfolios to find an image.
Tattoo recognition technology aims to achieve the same result as facial recognition. In fact, plans are in place to meld the two technologies – the FBI’s Next Generation Identification System calls for an automatic retrieval system for scars, marks and tattoos.
The use of prisoners’ tattoos in the Tatt-C project was highly controversial, as they were not freely given the ability to consent to the use of their personal identifying information. In several cases, the tattoos had names and birthdates, or were located in intimate places.
The Electronic Frontier Foundation (EFF), a civil rights advocacy group that broke this story, has called for an end to the use of prisoners’ body art to create a tattoo-matching database. The EFF, which calls tattoos “unique biometric identifiers,” noted that the NIST project has expanded to 100,000 images gathered by Florida’s Pinellas County Sheriff’s Office, the Michigan State Police and the Tennessee Department of Correction.
With increased tattoo recognition, people could be tracked or identified while walking down the street based solely on their body art.
Private company Face Forensics announced in January 2016 that its system can do just that – it can identify scars, marks and tattoos. Once a match is made, it then displays the face and name of the owner from a database. This is all done in a matter of seconds, allowing law enforcement officers to match a person to an image of their tattoo. Of course, as with any identification system, tattoo recognition is subject to errors.
According to a January 10, 2017 online news report, the National Institute of Standards and Technology stated, “The goal of the NIST project is to help ensure tattoo matching technologies are evaluated using sound science to improve accuracy and minimize mismatches.”
For now the focus is on tattoos from arrestees and prisoners, but technological advances potentially allow authorities to include photos taken from surveillance videos. The technology is being touted as a way to not only identify criminals, but also victims of crime or natural disasters. Yet tattoo identification has already been used for more questionable purposes.
California’s prison system has long used prisoners’ ink to classify them as gang members or associates and add them to gang databases. Tattoos have also been used by the U.S. State Department to deny visas to applicants because it was believed they had gang-related tattoos. And in one case in Tennessee, state prison officials mistakenly identified a prisoner’s military patch tattoo as a gang symbol.
“The number of people getting tattoos is rapidly growing. About 20 percent of the population has at least one tattoo, and this percentage is even higher [among] delinquents,” said Anil Jain, a Professor of Computer Science and Engineering at Michigan State University. “This system has huge implications for helping law enforcement with suspect and victim identification.”
The larger issue, however, is the impact of tattoo recognition on civil rights by using people’s body art to associate them with religious, political, criminal or other groups, and to create databases for use – or misuse – by law enforcement agencies.
Sources: Electronic Frontier Foundation, www.sciencecodex.com, www.fusion.net, www.hdiac.org, www.constitutionproject.org, www.oneworldidentity.com