IT Briefcase Exclusive Interview: The Value and Incresed Use of Facial Recognition Software

In this interview, we talk to Brian Strock, Manager, Public Safety Group at Numerica Corporation, about facial recognition software, the increased use of this technology by law enforcement, and Numerica’s latest software solution, Lumen FR.

• Q. Generally, how does facial recognition software work?

A. Facial recognition software uses advanced, deep learning algorithms to compare a “probe” photo, such as an image obtained from surveillance footage, against a database of known individuals to quickly identify any potential matches. The algorithms analyze the structure and characteristics of the face in the probe photo to compare against the previously indexed facial information in the database.

• Q. It seems like facial recognition is being used for everything from unlocking your phone to checking into the airport. How does law enforcement typically use facial recognition software?

A. The typical law enforcement use for facial recognition is to identify an unknown suspect in a criminal incident. A lawfully obtained probe image of an individual, such as a suspect in surveillance footage or even a composite sketch, is used to compare against a database of booking photos to attempt to make a positive identification. Facial recognition could also be used to assist in the identification of a person who lacks the capacity to identify him- or herself, such as a deceased or incapacitated person.

• Q. Recently, consumers have complained about retail stores using facial recognition software to track purchasing patterns. Are there similar concerns that law enforcement could use facial recognition software to track the behavior of citizens, regardless of whether or not they are involved in criminal activities?

A. Law enforcement uses facial recognition in a more active, rather than passive fashion. This means the technology is not used for general surveillance of the public without their knowledge or consent.

• Q. What kinds of safeguards are in place to make certain law enforcement uses facial recognition appropriately?

A. Agencies institute and enforce clear and specific policies on the uses of facial recognition technology to ensure there are no unlawful uses. Any potential matches from a facial recognition search are not considered positive identification and do not establish probable cause, without further investigation; rather, they are advisory in nature as an investigative lead only. Any possible connection or involvement of an individual to a criminal investigation must be determined through further analysis and investigation, just as any other lead would be vetted.

• Q. Numerica recently launched its own facial recognition software, Lumen FR. How does Lumen FR differ from previous facial recognition systems?

A. Lumen FR is the first and only integrate facial recognition technology, meaning it combines advanced facial recognition algorithms with a full set of known offender information from multiple sources (including criminal histories, vehicle and person links, and other criminal records information). Additionally, Lumen FR requires little to no manipulation or editing of the probe photo, whereas any previous facial recognition tools require a great deal of photo editing in complicated interfaces to ensure the photo is entirely front-facing or a certain quality to work with their algorithms. Lumen FR achieves matches even when the probe photo has poor lighting, off angles, or other photo quality issues that can hamper positive identification.

• Q. What are the advantages of law enforcement using Lumen FR over other facial recognition systems?

A. The primary advantage of Lumen FR over other systems is the integration with a host of other information sources with known offender information, coupled with the advanced recognition algorithms.  Lumen FR’s sophisticated analytical capabilities for automatically assimilating offender information is capable of rapidly identifying a small number of suspects matching a probe photo and other known details about a subject, such as vehicle associations or tattoo descriptions, then quickly and easily assessing those matches, resulting in better quality leads.

• Q. Does Lumen FR work with the other components of the Lumen software suite?

Yes, that’s what makes Lumen FR much more useful than a standalone facial recognition solution. It integrates directly with the Lumen Desktop and Lumen Mobile solutions for use both at a desktop workstation and in the field on a mobile device. The integration and simplicity make it a valuable tool for any user in a law enforcement agency who may need to quickly and easily identify potential suspects and generate investigative leads.

Brian Strock is Manager, Public Safety Group at Numerica Corporation. Numerica delivers state-of-the-art defense and law enforcement solutions to both government and industry customers. For its law enforcement customers, Numerica provides mission-critical insights to law enforcement, intelligence, and security professionals through the development of intuitive software solutions.