Clear75° WeatherClear75° Weather

Facial recognition technology proving effective for Pinellas deputies

Pinellas Deputy Mark Eastty, who got his mobile facial-recognition unit in May 2008, has “used it over 100 times now.” He said, “I’ve actually never had anyone give me grief over it.”

MELISSA LYTTLE | Times

Pinellas Deputy Mark Eastty, who got his mobile facial-recognition unit in May 2008, has “used it over 100 times now.” He said, “I’ve actually never had anyone give me grief over it.”

The man said his name was Erik Lopez and his birthday was Nov. 22, 1971, but Deputy Mark Eastty wasn't so sure. The man, a passenger in a car Eastty pulled over in February for expired tags, said he didn't have a driver's license. So the Pinellas deputy took the man's picture. He put the digital image into his laptop and ran it through software that sifted through 7.5 million jail mug shots. Within seconds, he had a hit. The man, it turned out, was Elisur Velazquez-Lopez of Clearwater. He was born in 1973. And he had an active warrant for failing to appear in court for a charge of solicitation of prostitution. Eastty took the man to jail. "If we had played the name game I could've been there 30 minutes," Eastty said. "If I take their picture, it'll take two."

Eight years ago, people scoffed when the Pinellas County Sheriff's Office applied for a $3.5 million federal grant to get a facial recognition system.

Critics slammed the technology as too new, too costly and ineffective. Similar facial recognition systems, which had debuted in airports and at the 2001 Super Bowl, had failed to produce any arrests or single out any terrorists. The Tampa Police Department chucked a program used to scan faces in Ybor City.

"We had no reason to believe that facial recognition would work," said Bruce Howie, chair of the legal panel for the Pinellas chapter of the American Civil Liberties Union, which decried the technology as a privacy invasion. "We always have to be concerned about the erosion of Fourth Amendment rights when new technology is applied."

But these days, the controversy has died down, the technology has improved, and facial recognition has become a seamless part of everyday work at the Sheriff's Office.

In fact, the facial recognition system has led to 463 arrests and nearly as many positive identifications that didn't result in arrests, said Scott McCallum, a system analyst who maintains the sheriff's system.

"In 2001 we pioneered this technology," McCallum said. "It is now becoming a recognized tool because of its successes."

• • •

That doesn't mean there still aren't misconceptions.

McCallum, who has been involved with the project since day one, said people still view facial recognition as a far-fetched piece of technology one might see on a TV show.

But in Pinellas, it's used daily in the courthouse and the jail, and by 170 patrol deputies like Eastty who can access the technology while on the road.

In every application, the technology works the same way.

First, a digital image of a person's face is entered into the system, which maps out 128 facial landmark points and turns the image into an algorithm.

From there, the software compares the algorithm against others in the database. It is oblivious to things like a person's hairstyle, gender, race or age, McCallum said.

Seconds later, a batch of possible matches shows up, which will include other data like a birth date and Social Security number.

It's up to the human user, however, to make the final call.

And, if for instance, the system doesn't produce a match or proves that the person is being truthful about his or her identity, that photo is not kept in the database.

Deputies using the system on the road also must get consent before taking someone's photograph, McCallum said.

"I've used it over 100 times now," said Eastty, who got his mobile unit in May 2008. "I've actually never had anyone give me grief over it."

• • •

Pinellas paid for the technology with three federal grants from 2001 to 2004 that totaled nearly $8 million.

The only cost associated with it now is a $148,500 yearly maintenance contract. But even that amount is cut in half thanks to leftover grant money, said sheriff's spokeswoman Cecilia Barreda.

The Miami-Dade Police Department is the only other law enforcement agency in Florida with its own facial recognition system, Barreda said. Elsewhere in the country, a handful of other law enforcement agencies use it, Barreda said, "but probably none with the volume and data as us."

"This is the largest system in the country available for law enforcement," McCallum said.

So far, 7.5 million mug shots dating back to 1994 are in the database, drawing on photos from 15 partner sheriff's offices around the state and the Department of Corrections.

"The goal when we first started was just to enroll our own data," McCallum said. "It's exceeded what we thought."

People are starting to notice.

McCallum regularly gives presentations about how the technology has helped in the jail booking process, especially with uncooperative inmates or John Does. There's a waiting list of deputies who want to get the technology in their patrol cars. And a police department in Fairfax, Va., which has a similar booking system, has been in talks with Pinellas about a partnership.

Pinellas also gets regular requests from law enforcement agencies in and out of state to perform facial searches.

The Pasco County Sheriff's Office, which recently added its jail mug shots to Pinellas' database, had its first success with the system a few weeks ago. And the Hillsborough County Sheriff's Office, whose mug shots are also in the database, says it has been paying attention to the success Pinellas has been having.

"I'm excited about it," said Pasco Lt. Barbara Taylor. "With any new thing you have to weigh the pros and cons, and I really haven't seen any cons."

Critics admit that they haven't kept tabs on the facial recognition system through the years, but local ACLU members say that as long as authorities are using the technology in a limited manner — as Pinellas says it is — then the privacy concerns are allayed.

"If it has a high success rate, then maybe it is an effective tool," Howie said.

Kameel Stanley can be reached at (727) 893-8643 or kstanley@sptimes.com.

Facial recognition success stories

A lot has changed since the Pinellas County Sheriff's Office decided to jump into the facial recognition game in 2001. Once thought to be invasive and risky, the technology has proved to be an effective tool law enforcement can use to help verify and identify people. Here are a few successes.

Shannon Renee Horton: On Sept. 11, 2004, a couple of months after Pinellas put facial recognition software in some of its patrol cars, deputies were called to a mobile home park to contend with a disorderly woman. When they got there, they found Horton, who told deputies over and over that her name was Debra Lynn Smith. She claimed to be from Tennessee and had no identification on her. After she consented to the deputies taking her photo, officials got four matches from Horton's previous arrests in Pinellas County. They also found that she had two warrants for violating probation on prostitution and cocaine charges.

George Angel Meulener: On May 7, the North Miami Police Department called the Sheriff's Office seeking help identifying a bank robber whose face had been captured in a high-quality surveillance photo. The facial recognition system provided the name and photo of Meulener, who witnesses later picked out of a photo lineup. Later that day, police arrested Meulener, who they said confessed.

Carlos Mayolino: On July 1, the Sheriff's Office got a request from the Miami Police Department, which was looking for another bank robbery suspect. Miami police had gotten an alleged ID card with a picture of the suspect, but the name listed was Juan Saldana. Pinellas compared the photo to those in its database and found a match for Mayolino, who had been arrested in 2003 in Orange County. The next morning, Mayolino was arrested in St. Louis.

Facial recognition technology proving effective for Pinellas deputies 07/17/09 [Last modified: Thursday, July 23, 2009 5:59pm]

© 2014 Tampa Bay Times

    

Join the discussion: Click to view comments, add yours

Loading...