OHLEG’s facial recognition system helped Akron police detectives Jim Pasheilich and Aron Hanlon identify a suspect within hours of a January murder and robbery. The next day, they had the man in custody and secured information on two other suspects.
Marshall Jones was riding in an SUV with three other men when he was robbed, shot, and pushed onto a city street. Jones’ girlfriend said he had left their home a little while earlier with a man she knew only as C.J. She had a photo that C.J. failed to delete from a computer he sold the couple, and she showed it to Pasheilich. The detective took a picture of the image on the screen and ran it through the facial recognition system on the Ohio Law Enforcement Gateway (OHLEG).
Within the search results, an Ohio inmate photo of Charles Fortson had the top probability score. Ironically, it was the same photo that Fortson had failed to delete from his old computer and that Pasheilich used to search the database. Officers found and questioned Fortson the next day and, after investigating, charged him in the case. He pleaded guilty June 5 to manslaughter, aggravated robbery, and a gun specification and agreed to testify against the shooter.
“It was the first case I’ve used it in, and it was hugely beneficial,” Pasheilich said of the facial recognition software, which the Attorney General’s Office made available to Ohio law enforcement in 2013. “In
How the software works
The Ohio Organized Crime Investigations Commission piloted a facial recognition program in 2008, and federal grants over the next two years helped build a photo repository to make such a system available to all Ohio law enforcement agencies. The Bureau of Criminal Investigation (BCI) now maintains facial recognition technology on OHLEG.
The software draws from a database of about 24 million photos from the Ohio Department of Rehabilitation and Correction, sex offender registry, Bureau of Motor Vehicles, Ohio Courts Network, and various law enforcement agencies.
Manufactured by 3M Cogent Inc., the program uses algorithms to find and isolate numerical information in faces, such as the relative position, size, and shape of the eyes, nose, cheekbones, and jaw. To conduct a facial recognition search, an investigator uploads a probe image of an unknown person; the system compares it with known images in the database and returns a probability score indicating the likelihood
of a match. this case, it was very helpful. It sped up the process of identifying and locating a suspect.”
An innovative resource among many
Facial recognition is a valuable tool that can supplement other investigative strategies, said Joe Dietz, BCI’s director of investigative services. It is among eight search types and 17 data sources available to law enforcement within the OHLEG search engine, which is one of 15 OHLEG components.
“If investigators have a photo of a suspect but not his or her name, facial recognition may give them the lead they need,” he said. “No law enforcement officer would rely solely on a facial recognition search result to identify a suspect, but it can provide possible suspects who require further investigation.”
Since the system became available a year ago, 547 Ohio agencies have used it to conduct about 8,500 searches of the photo database. By comparison, officers have conducted more than 3.6 million other types of searches within OHLEG-SE.
“Facial recognition still accounts for a very small percentage of Ohio law enforcement agencies’ OHLEG searches,” Dietz said, “but that will increase as more officers learn about the system and how it can further certain investigations.”
Let us know what you think
BCI is interested in law enforcement officers’ experiences with OHLEG’s facial recognition system. If you have questions, successes, or other feedback, please email
OHLEGSupport@OhioAttorneyGeneral.gov.
OHLEG upgrades
The Attorney General’s Office continues to roll out improvements to OHLEG following an advisory group’s recommendations in 2013.
“Our goal is to preserve the access to and integrity of this tool while respecting the privacy interests of the citizens of Ohio,” said BCI’s Joe Dietz, who oversees OHLEG.
Some recent updates:
-
Use of the facial recognition system has been restricted to Ohio law enforcement officers designated by their agency CEOs to require access to the software. As of May 31, approximately 5,100 of OHLEG’s 23,634 users had access to the facial recognition function.
-
Agency CEOs have the ability to immediately terminate staff members’ OHLEG access.
-
An eight-minute training video on OHLEG highlights the information system’s tools and spells out best practices and expectations for use as well as consequences for misuse. Under Ohio Revised Code Section 2913.04, using OHLEG for anything other than criminal justice purposes is a fifth-degree felony.
-
Two auditors have been hired to conduct random and directed audits of OHLEG use. They are part of a new OHLEG Quality Assurance Unit, which also includes two trainers.
-
Users must update their passwords every 90 days to maintain OHLEG access.
-
Updated rules and regulations and a data use policy have been implemented and posted to the website.
For more information: View the training resources at www.OHLEG.org or contact the OHLEG Help Desk at
OHLEGSupport@OhioAttorneyGeneral.gov or 866-406-4534.