MPR News Chief Meteorologist Paul Huttner signs off

MPR News helps you turn down the noise and build shared understanding. Turn up your support for this public resource and keep trusted journalism accessible to all. 

Turn Up Your Support

Walter Yovany-Gomez evaded authorities for years before the FBI put him on its Ten Most Wanted Fugitives list.

Gomez, a member of the MS-13 street gang, was wanted in connection with a brutal murder in Plainfield, N.J., that took place in May 2011. Police almost nabbed him a month afterward — but Gomez jumped out a second-story window and escaped.

Investigators finally tracked him down and arrested him in August 2017 in a gym parking lot in Northern Virginia.

Gomez's capture made headlines at the time, but the details of how investigators put the pieces together — with the help of a tipster and facial recognition technology — have not been previously reported.

Interviews by NPR with law enforcement officials and others now reveal the role that digital facial recognition software played in the case at a time when authorities' use of such technology is under increasing public scrutiny.


Some lawmakers and civil liberties advocates warn that there are little to no transparency and few rules governing the use of facial scan software and its vast surveillance potential, leaving the door open to possible abuse. Researchers, meanwhile, have found the software is less accurate with women and 

The FBI began phasing in its use of facial analysis software around 2011.

Local and state police also have used the technology for years, and proponents say it's an important investigative tool that can help find missing children, prevent driver's license fraud — or, in Gomez's case, help track down a suspected killer.

Gomez, a Honduran national who was in the U.S. illegally, belonged to a branch of MS-13 in Plainfield, outside New York City.

The gang began in Los Angeles in the 1980s but has expanded since then and now has a presence across the U.S. and Central America. Like many gangs, MS-13 is involved in drug dealing, prostitution, extortion and murder. It tends to focus many of its activities in immigrant communities and has set itself apart through its sheer brutality.

MS-13 is not a rigidly hierarchical organization, according to experts. Instead, it has various branches, or cliques, that operate under the group's umbrella but with their own internal leadership and rules.

In the spring of 2011, the Plainfield clique's leadership ordered Gomez and another man, Cruz Flores, to kill a potential recruit, Julio Matute, who was suspected of disrespecting MS-13 because he associated with members of a rival gang.

On May 8, 2011, Gomez and Flores spent an evening at Matute's apartment drinking, smoking weed and watching TV, according to investigators. They then beat him in the head with a baseball bat, stabbed him 17 times in the back with a screwdriver and slit his throat.

"He was stabbed so many times that when his body was discovered a week later, police officers thought that he had been shot with a shotgun," said FBI Special Agent Dan Brunner of the bureau's Newark Division.

Two years later, federal prosecutors indicted more than a dozen members of the Plainfield MS-13 clique for racketeering, murder and other crimes. Gomez and Flores were charged in that indictment in connection with Matute's murder.

By the end of 2016, Flores and the others had been apprehended, tried and convicted.

The one member who remained at large was Gomez.


In April 2017, the FBI placed him on its 

. His official poster included three photographs — a young man with short black hair and an earring — and described him as an extremely violent criminal. It offered a $100,000 reward for information that led to his arrest.

A few months after Gomez appeared on the most wanted list, the FBI received a tip with a possible lead. The information made its way to the FBI's Washington Field Office and the desk of Special Agent Richard Stallings.

Investigators believed the tipster appeared legitimate; he knew the players and gang dynamics in Northern Virginia. What the tipster told them was that he recognized the man in Gomez's most wanted poster. He knew him by a different name, but he was sure it was the same man.

After a few more meetings, the source provided more information to help push the investigation forward, including Facebook profiles that he said had pictures of the individual he believed to be Gomez.

"He gave us some screen names we were able to track down," Stallings said.

Stallings and his fellow investigators went through the Facebook pages and downloaded photos that the source identified as Gomez.

Not any photo will do; the perspective has to be right and the file itself must be good enough to be of use to investigators.

"The quality of the picture has to be reviewable. If it doesn't have enough of the facial features, they're not going to be able to plug it in and have a solid match," Stallings said. "So we had to screen and find enough of the pictures that we thought had the qualities needed to make the matches and assessments."


Stallings sent those photos off to the FBI's 

, which ran the photos against those contained in the bureau's databases.


"We sent this information off to say, 'Hey, can we identify this guy? Are we looking at the right guy?' " Stallings said.

The FACE Services Unit came back with a match — but not for Gomez. Instead, the photos matched a man in official records named Jesus Lopez Centoreo.

Centoreo had been picked up for marijuana possession in 2014 after jumping a Metro turnstile in Arlington, Va. He had gone through initial processing — his fingerprints were taken and he was photographed for a mug shot — but was then released. He didn't show up for his court date, and so a warrant had been issued for his arrest.

Despite the name confusion, investigators still felt their source was solid, Stallings said, in part because there were other physical identifiers, including tattoos, that made them think they were on the right track.

They also had another lead to pursue from the Facebook pages.

Agents had found several photos of the suspect with a woman. They ran photos of her through the bureau's facial recognition software.


They got a match from the criminal mug shot database. With the woman's name they were able to get an address, and the FBI and officers from the 

 gang task force used that to begin keeping tabs on her.

On Aug. 12, 2017, officers from the task force called Stallings and said they were heading over to stake out the woman's residence.

"We were hoping to get lucky, get a vehicle, get something else to put a building block together with," Stallings said. "And so they went down there sitting on it and a couple of hours later, I get this call back from their sergeant: 'We got our guy.' "

The surveillance team from the Fairfax County gang unit had followed the woman to a gym parking lot in Woodbridge, Va. A man who looked like Gomez walked up to the woman's car.

Task force members swooped in and made an arrest. The name the man gave was Jesus Lopez Centoreo — which matched the fingerprints on file.

"But we knew from other physical identifiers, from his tattoos, from the picture that we had of him that we had the right guy," Stallings said.

He says the officers showed the man the most wanted poster that named Gomez and asked whether it was him. He said yes.

"So he self-identified as Gomez the night we arrested him," Stallings said. "I think he just knew that it was done. He knew that we hadn't made a mistake in identifying him and putting our hands on him and making the arrest. We had done our corroboration."

Gomez was transferred to New Jersey, where he reached a deal with the government in February. He pleaded guilty to one count of racketeering conspiracy. He admitted in his criminal information to Matute's murder.


Last month, a federal judge sentenced him to 25 years in prison.


Crime, Law and Justice

Technology

Ryan Lucas

Author

Even as civil libertarians warn about the increased use of facial recognition by federal authorities, the bureau is using it to track down some of its most wanted suspects.

How a tip — and facial recognition technology — helped the FBI catch a killer

Go Deeper.

Like this?

News you can use in your inbox