Quantcast
Channel: Crime-public_Safety Rss Full Text Mobile
Viewing all articles
Browse latest Browse all 2542

Police using facial recognition software

$
0
0

DAYTONA BEACH - Being a 50-year-old white man didn’t make James Morrison unique in Florida.

But his crooked mouth, the freckle on his forehead and the space between his eyes did, and those characteristics led to him being positively identified through facial recognition software as a suspect in a bank fraud case.

Morrison’s identification is one of two cases this year where Daytona Beach police were able to use the Face Analysis Comparison Examination System, or FACES, to identify the suspect of a crime.

“This is not what you see on Criminal Minds or CSI or NCIS,” said Jackie Flory, the specialist for the department’s Criminal Investigations Division.

The Volusia County Sheriff’s Office and the Daytona Beach Police Department were the first local agencies to start using the system. Other law enforcement agencies — police in the cities of Ormond Beach, South Daytona, Daytona Beach Shores, Edgewater, Orange City and the Flagler County Sheriff’s Office — also have access to FACES.

When detectives bring Flory a photo, often a still from surveillance video, she puts it into the system, which was developed by the Pinellas County Sheriff’s Office in 2001, and narrows down the search parameters — based on approximate age and other characteristics — to ensure she gets the best results.

The FACES database is composed of photos from about 40 state and federal agencies including jail booking photos and Florida driver’s license pictures.

Cecilia Barreda, spokeswoman for the Pinellas County Sheriff’s Office, said the program then sorts through a plethora of pictures, nearly 30 million, scanning faces to find the best matches based on the facial value assigned to the photo in question. The facial value is determined based on points on the face.

Next, the system returns with a gallery of results, ranked by which photo is the closest potential match.

Barreda said it’s then up to the user to determine if any of the results is a match.

“It’s not the end all, be all,” Barreda said. “It’s one more investigative tool that we have.”

She said the goal is to be able to put a name with an unknown face, which is what Flory was able to do in regard to Morrison.

Morrison is one of a handful of people, including a local Wells Fargo employee, suspected of being involved in a bank fraud scheme.

Jomar Lizardo, 30, worked at the bank at 1100 W. Granada Blvd. and was arrested Jan. 31 after Ormond Beach police were able to connect him to the case, which they started investigating in November. Lizardo is facing charges of grand theft more than $20,000, organized scheme to defraud more than $20,000, conspiracy organized scheme to defraud and using a two-way communication device to commit a felony, records show. So far, Lizardo is the only person charged in the case.

Ormond Beach detectives first brought Flory the Wells Fargo surveillance photo of Morrison, who police say was attempting to withdraw $4,650 from an Arizona man’s account Jan. 14, but Flory was unable to get any hits because of the quality and angle of the photo.

Flory said more often than not banks and other businesses with surveillance cameras don’t have them set up to capture straight-on photos, which work best for FACES.

“We can try to tilt the photo, but if they’re looking down with a ball cap and sunglasses on, there’s no way,” Flory said.

She then remembered the Ormond Beach police report said that Morrison had left the fake Arizona ID behind when he fled the bank.

Flory said she was able to borrow the ID from Ormond Beach police and scan the photo into the system, ultimately finding a match.

But officials say making a positive identification won’t always be the case.

When Ormond Beach police sent a photo of a female suspect in the same bank fraud case as Morrison to the Volusia County Sheriff’s Office to run through FACES, investigators were unable to find a match because the bank teller’s window distorted the suspect’s face in the surveillance still, according to an Ormond Beach police report.

“FACES is a great program, and it’s a tool, but it’s not perfect,” Volusia County sheriff’s Lt. Jim Morgan said.

Morgan said a positive identification made by a real person, without the help of a computer program, is preferred, but FACES is still helpful.

While law enforcement officials know what the facial recognition system is capable of, officials say the public’s perception of its capabilities can be distorted.

“We do not have facial recognition cameras throughout the city,” Flory said. “People shouldn’t be afraid of this. We’re not out to spy on people. We’re trying to catch the bad guys. We’re not out to get the good guy.”

Jon Mills, privacy expert and dean emeritus of the University of Florida’s Levin College of Law, said he believes the system is mostly reliable, but he could see why some of the public has a problem with the program.

“I think a lot of the tech concerns are warranted, but that picture is probably public record, and if it’s public record then you don’t have a reasonable expectation of privacy,” said Mills, a professor of law.

Bruce Jacob, a constitutional law expert and dean emeritus of Stetson University’s College of Law, questioned how people may feel about their photo being a part of database used and maintained by law enforcement.

“Whether it violates the Fourth Amendment, I just don’t know,” Jacob said.

Barry Butin, a criminal defense lawyer and co-legal panel Chair of the American Civil Liberties Union in Broward County, said he could understand why Florida residents may take issue with their driver’s license photos — which aren’t public record — being a part of the system, but he doesn’t see FACES violating any constitutional rights.

“It’s a different situation when you’re committing a crime,” Butin said. “I guess it’s not great for privacy, but there’s two sides to every coin.”
 


Viewing all articles
Browse latest Browse all 2542

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>