FBI seeks industry input on biometric algorithms, AI-driven tattoo recognition

The Federal Bureau of Investigation (FBI) has issued a new Request for Information (RFI) seeking industry input on the next generation of biometric matching algorithms, signaling growing federal interest in both traditional and emerging identification technologies, as well as reviving AI-enabled tattoo recognition.

The move indicates that the FBI is positioning itself for a new phase of biometric modernization, one that blends traditional identifiers with emerging AI-based tools that could reshape how U.S. law enforcement processes identity data.

The RFI is issued by the bureau’s Criminal Justice Information Services (CJIS) Division and emphasized that the announcement is strictly for market research and not a solicitation for contracts. CJIS manages the massive Next Generation Identification (NGI) system, the world’s largest electronic repository of biometrics and criminal history data.

The division is now surveying the commercial landscape for advanced algorithms across four categories: ten print fingerprints, latent fingerprints, facial recognition, iris recognition, and tattoos.

The FBI said it wants to understand the current capabilities of industry developed systems to guide future decisions about enhancing NGI’s identification tools.

Under federal law, the Attorney General is authorized to collect and maintain biometric and identification records for criminal justice purposes, but the notice reiterates longstanding limits on the bureau’s powers.

The FBI noted that it is barred from engaging in investigative activity solely on the basis of First Amendment protected behavior and from maintaining records about such activity unless tied to an authorized investigation.

To evaluate new technologies, the FBI is asking applicable vendors to submit their algorithms to the National Institute of Standards and Technology (NIST), which evaluates fingerprints, faces, irises, and, increasingly, other biometric markers to help determine how accurate, reliable, and scalable commercial algorithms are before they are considered for law enforcement use.

The FBI also is exploring a newer frontier in automated, image-based tattoo recognition. The NGI system today supports only text-based searches for tattoos, relying on descriptions provided at the time of arrest.

Advances in AI though have enabled more precise image-matching techniques that could allow investigators to search for tattoo images directly rather than depend on sometimes inconsistent verbal descriptions.

The FBI said such capabilities could support both criminal investigations and humanitarian missions, such as identifying victims.

In 2014, NIST began the FBI-backed Tattoo Recognition Technology Evaluation (Tatt-E) program to benchmark recognition accuracy and identify limitations of these emerging systems to perform automated image-based tattoo recognition. A public report on Tatt-E was published in 2018.

In February 2016, NIST published Tattoo Recognition Technology Best Practices.

That same year, an Electronic Frontier Foundation (EFF) investigation reported finding “that NIST had skipped over key ethical oversight processes and privacy protections with its earlier experiments called Tatt-C, which is short for the Tattoo Recognition Challenge.”

EFF said in November 2018 that the “experiment promoted using tattoo recognition technology to investigate people’s beliefs and memberships, including their religion. The more recent Tatt-E, however, did not test for ‘tattoo similarity’ – the ability to match tattoos that are similar in theme in design, but belong to different people.”

Continuing, EFF stated that, “A database of images captured from incarcerated people was provided to third parties – including private corporations and academic institutions – with little regard for the privacy implications.”

“After EFF called out NIST,” EFF said, “the agency retroactively altered its presentations and reports, including eliminating problematic information and replacing images of inmate tattoos in a ‘Best Practices’ poster with topless photos of a researcher with marker drawn all over his body. The agency also pledged to implement new oversight procedures.”

Last year, the European Commission’s Joint Research Centre, along with EUROPOL, released the paper, Tattoo Recognition with Artificial Data, which concluded that tattoo recognition – especially from low-quality, real-world video – remains a fundamentally difficult problem, even when using modern AI techniques.

The proficiency test using synthetic 3D-avatar–generated datasets revealed that while some methods (particularly keypoint-based systems and CLIP-based approaches) showed promise, no single algorithm consistently performed well across all probe conditions.

Overall, the paper concludes that synthetic datasets are useful and capable of exposing real-world weaknesses, but substantial research is required before automated tattoo recognition can reliably support forensic casework.

Notably, the EU report, and NIST’s Tattoo Recognition Technology Best Practices, both showed that tattoo recognition systems – whether for law enforcement booking stations or forensic video analysis – can only succeed when the input images meet minimum quality standards.

Both documents show that current AI-based tattoo recognition breaks down most often because of capture artifacts, not algorithmic deficiencies alone.

Still, the FBI said in its RFI that “image-based tattoo searching could provide a more objective and efficient search capability in support of law enforcement investigations and humanitarian efforts. As a result, the FBI is partnering with NIST to enable benchmark testing of industry available image-based tattoo matching algorithms.”

The RFI indicates that NIST is continuing to conduct “the Tattoo Recognition Technology Evaluation to assess the capabilities of image-based tattoo matching algorithms, with objectives of measuring recognition accuracy and understanding technology limitations.”

In August, NIST issued issued an Application and Agreement to Participate in the Tattoo Recognition Technology-Evaluation.

This year, NIST also released a draft Evaluation of Tattoo Recognition Algorithms Concept, Evaluation Plan, and API Version 0.1.

Details of the evaluation, including schedule, participation instructions, and application programming interface (API) can be reviewed here.

The API is currently open for public comment until December 31. Comments and questions may be submitted to tatt-e@nist.gov.

Vendors interested in being considered in FBI market research and ultimately supplying the FBI with automated biometric matching algorithms, are encouraged to submit their algorithms to NIST’s technology evaluation.

Vendors are not required to submit an algorithm for all modalities, and each one submitted for a given category will be considered.

The bureau stressed that formal proposals are neither required nor accepted at this stage.

The FBI said it “intends to release subsequent RFIs to industry to further refine our analysis of the market for these technologies. The RFIs may include refined benchmarks for accuracy and performance, outline requirements for compatibility and use within the existing environment, history of performance, price points and packaging, and other areas of focus.”

Article Topics

biometric matching  |  biometrics  |  criminal ID  |  FBI  |  forensics  |  law enforcement  |  NGI  |  RFI  |  tattoo recognition