More Than 25 Years of Independent Biometric Performance Testing

First place. In the first Fingerprint Verification Competition, Neurotechnology's VeriFinger algorithm achieved the best reliability results among all participants. This marked the beginning of the company's participation in independently organized fingerprint algorithm competitions.

Neurotechnology participated in four editions of the Fingerprint Verification Competition under its former name, Neurotechnologija. The company adopted the name Neurotechnology in April 2008.

Organized by the University of Bologna's Biometric System Laboratory together with other academic biometric research institutions, the competitions evaluated fingerprint verification algorithms using standardized datasets and several measures of recognition accuracy.

Neurotechnology continued its participation in FVC2002, where its fingerprint recognition algorithm received one silver and two bronze medals across the evaluated datasets and performance measures.

Neurotechnology's fingerprint evaluation history with NIST dates back to 2003, when the company participated in the Fingerprint Vendor Technology Evaluation (FpVTE) under its former name, Neurotechnologija.

The algorithm showed one of the strongest reliability results in the Middle Scale Test. FpVTE assessed one-to-many fingerprint identification and provided an early independent measure of the technology's suitability for larger biometric databases.

In FVC2004, Neurotechnology participated with algorithms designed for both general-purpose and resource-constrained applications. VeriFinger received four gold, three silver and two bronze medals in the Open Category. FingerCell, the company's fingerprint technology for embedded and lower-resource platforms, received one gold, six silver and three bronze medals in the Light Category.

The separate categories demonstrated the company's ability to develop fingerprint algorithms for different operational requirements, from full-scale software systems to platforms with more limited computing resources.

First place. In FVC2006, Neurotechnology's P058 algorithm achieved the highest overall ranking in the Open Category according to Average Zero FMR, a measure of the false rejection rate when no false matches are permitted. The algorithm received four gold, two silver and two bronze medals in this category.

A separate Neurotechnology algorithm ranked second in the Light Category according to Average Zero FMR and received one gold and four bronze medals.

Zero FMR is particularly relevant to high-security biometric applications because it measures recognition performance at an operating point where false acceptance is reduced to zero in the evaluated dataset. At this threshold, stronger algorithms reject fewer legitimate users while maintaining the strict false-match requirement.

Together, the four FVC editions established an early independent performance record for Neurotechnology's fingerprint recognition technology, covering both recognition reliability and performance across different types of fingerprint images and computing environments.

The company has also participated in the Minutiae Interoperability Exchange (MINEX) evaluations for many years. Neurotechnology was recognized as MINEX compliant in 2007 for both fingerprint template generation and matching.

MINEX focuses on the interoperability of fingerprint templates created and matched by different implementations. This is particularly relevant to systems in which biometric records must remain usable across equipment, software versions or organizations.

In 2011, Neurotechnology's WSQ 3.1 implementation received seven FBI certifications, one for each supported computing platform.

Neurotechnology placed second in the Ongoing MINEX ranking in 2014.

Neurotechnology's participation continued in FpVTE 2012, where algorithms based on MegaMatcher SDK and MegaMatcher Accelerator were evaluated in one-to-many fingerprint identification scenarios involving single fingers, flat tenprints and more complex tenprint records.

When NIST published the final report in 2015, the company's submission ranked fourth for overall accuracy. In Class C, it was four times more accurate than the second-fastest algorithm in a test involving 5 million tenprints and a 90-second search limit.

Neurotechnology began participating in MINEX III in 2015.

MINEX III remains active, meaning that the results may change as participants submit updated algorithms.

Its MegaMatcher On Card algorithm passed MINEX III testing in 2016, followed by further submissions for template generation and matching.

MINEX III is the ongoing evaluation of fingerprint template interoperability for the US PIV program.

First for interoperability. In 2017, MegaMatcher SDK fingerprint technology ranked as the most interoperable matcher and the fourth most accurate native template matcher among the MINEX III-compliant matchers evaluated at the time.

Neurotechnology's first participation in the NIST Proprietary Fingerprint Template evaluations dates back to 2018, when the company submitted a fingerprint recognition algorithm to PFT II. Its latest PFT II submission, identified as 4E, ranked among the most accurate algorithms across all 33 experiments. It was also the second fastest during enrollment and the second or third fastest in fingerprint template matching.

PFT II evaluated one-to-one fingerprint verification using proprietary templates and datasets containing up to 120,000 subjects. The program ended in 2019, after which Neurotechnology continued its participation in PFT III.

First for interoperability. In 2018, its template generator ranked first for interoperability and its matcher ranked second.

First place. Neurotechnology entered PFT III in 2019 and took the top position on entry to the program.

The company has submitted successive fingerprint algorithm versions to measure one-to-one matching performance using proprietary templates.

First place. In 2019, a Neurotechnology submission achieved first place in the template generator category.

Active evaluation. Neurotechnology participated in NIST SlapSeg III, which evaluated algorithms that separate individual fingerprints from images containing several fingers captured simultaneously. The published results show tradeoffs between speed and accuracy; the company's submission was neither the fastest nor the most accurate across the four reported categories.

Slap segmentation is important in enrollment systems that capture four-finger slaps, including national ID, voter registration, border control and criminal registration workflows.

First place, active evaluation. In 2020, Neurotechnology's fingerprint extractor and matcher achieved the top result in the FV-STD-1.0 verification benchmark.

The FVC-onGoing platform provides continuing benchmarks for fingerprint verification, ISO-template matching, fingerprint indexing and related biometric tasks. Unlike the earlier FVC competitions, the platform accepts new algorithms on an ongoing basis while keeping the evaluation datasets fixed.

First place, active evaluation. In 2020, Neurotechnology's fingerprint extractor and matcher achieved the top result in the FV-HARD-1.0 verification benchmark. These tests evaluate one-to-one fingerprint matching, including more difficult image conditions.

First place. Neurotechnology held the top position in NIST PFT evaluations continuously since 2019, confirmed in successive PFT III result releases in 2022, 2023, 2024 and 2025.

In October 2023, smart-card submission 020A was confirmed as MINEX III compliant.

Separate smart-card-oriented submissions have also been evaluated for compact template generation and matching. Submission 020A, published in October 2023, is the company's latest smart-card submission and meets the applicable interoperability and accuracy requirements.

First place. Neurotechnology participated in the first Latent in the Wild Fingerprint Recognition Competition, held as part of the 2024 IEEE International Joint Conference on Biometrics. The competition evaluated algorithms using latent fingerprints collected from different surfaces under less controlled conditions. One Neurotechnology author contributed to the resulting competition publication.

VeriFinger 13.1 ranked first among three valid submissions from six registered teams. It achieved the highest area under the ROC curve at 0.854 and an equal error rate of 0.228. Its 6.2% failure-to-enroll rate represented a generation-over-generation improvement from the 40.8% rate of the earlier VeriFinger 12.3 baseline; it was not the best failure-to-enroll result in the competition. VeriFinger 13.1 also recorded the lowest computational cost among the submitted solutions, processing a fingerprint pair in 0.78 seconds.

The results demonstrated improvements over the earlier VeriFinger version in enrollment reliability and overall matching performance.

First place. Neurotechnology began its publicly documented participation in the NIST Evaluation of Latent Friction Ridge Technology in 2024. In December of that year, the company announced that its first publicly reported ELFT submission had achieved the highest accuracy across most of the evaluated datasets and ranked among the leading algorithms for feature extraction speed.

Latent friction ridge identification covers marks recovered from crime scenes, searched against ten-print and palm databases. ELFT measures performance in both identification and investigation scenarios. Identification tests determine whether a system can return the correct candidate within a specified list, while investigation tests assess candidate-ranking performance.

First place. In 2025, the Unique Identification Authority of India named Neurotechnology the winner of its Biometrics SDK Benchmarking Challenge for fingerprint recognition. The company's algorithm achieved first place, showing the strongest one-to-one fingerprint matching performance among the solutions selected for the final evaluation.

A notable feature of the challenge was its use of a unique longitudinal fingerprint dataset. The initial samples were collected from children aged 5–10, with corresponding samples captured again after a period of 5–10 years. This allowed UIDAI to evaluate age invariance: the ability of fingerprint verification algorithms to recognize the same individuals as they grow and their fingerprints change over time.

The competition attracted 2,106 applications. Following technical screening and strict security requirements, three submissions were selected for the final evaluation, with Neurotechnology securing first place.

First place, active evaluation. Neurotechnology participated in the Friction Ridge Image and Features Technology Evaluation Exemplar One-to-Many (FRIF TE E1N), the successor to FpVTE (2012) for large-scale fingerprint identification.

In 2025, its submission ranked first in the majority of NIST FRIF TE E1N test categories, including index-finger and multi-finger identification scenarios.

The Neurotechnology+0106 submission tied for first with one other participant in Class A index-finger identification. In Class B, it ranked second of 11 on individual slaps. The submission achieved zero errors in four experiments.

It ranked second in Class C ten-finger identification. Against 5 million records, median search times were 613 ms for plain ten-finger searches and 560 ms for rolled searches, with zero API timeouts. FRIF TE E1N remains active and evaluates one-to-many identification across different numbers and types of fingerprint images.

Neurotechnology continued its involvement in the second Latent in the Wild Fingerprint Recognition Competition, held as part of IJCB 2025. The second edition expanded the evaluation from a single recognition task to two separate tracks: latent fingerprint recognition and latent fingerprint quality assessment.

The recognition track evaluated the ability of submitted algorithms to compare latent fingerprints with corresponding reference fingerprints. The new quality assessment track examined whether algorithms could estimate the usability and recognition quality of latent fingerprint images. This is particularly relevant to forensic workflows, where image quality can affect both automated candidate searches and the work of fingerprint examiners.

Neurotechnology co-authored the publication presenting the competition and its results. The company's involvement in consecutive editions provided an opportunity to evaluate the continued development of its fingerprint technology on challenging latent fingerprint data.

[INTERNAL CONFIRMATION REQUIRED: Add Neurotechnology's exact 2025 recognition and quality-assessment placements after checking the complete IJCB paper. Until then, state participation only; the results are published, but the company's placement has not been confirmed from the paywalled paper.]

In 2025, Neurotechnology's slap fingerprint segmentation algorithm showed off as a top performer, featuring the fastest performance and almost the best accuracy in most categories of the evaluation.

[INTERNAL CONFIRMATION REQUIRED: This statement conflicts with the SlapSeg III entry above, which reports that the submission was neither the fastest nor the most accurate across the four reported categories. Confirm which result release each statement describes before publication.]

First place, active evaluation. In 2026, the company repeated a top-ranking achievement in ELFT, securing first positions in both identification and investigation scenarios on key datasets.

The submission ranked first in both scenarios on the IARPA N2N and FBI-Provided Solved #1 datasets. It also achieved leading positions in both scenarios on the DoD-Provided Dataset #1 and shared first place on the FBI Laboratory dataset.

First place, active evaluation. In 2026, the company continued to lead key MINEX III results. Its algorithms provided the most interoperable PIV Level 1 template generator, the most accurate native Level 2 generator and matcher pair, and the leading interoperable Level 1 generator and matcher pair.

First place, active evaluation. In the latest PFT III results reviewed in 2026, one Neurotechnology submission achieved the highest accuracy, while another emphasized faster matching and smaller templates, demonstrating the company's ability to refine algorithms for different operational priorities.

PFT III remains an active evaluation.

First place. In 2019, Neurotechnology achieved first place in the FVC-onGoing palmprint recognition benchmark.

Neurotechnology's participation in NIST face recognition evaluations includes both earlier and current programs. In 2015, the company submitted an experimental face recognition engine integrated with a face tracker to the Face in Video Evaluation (FIVE), which assessed face recognition of non-cooperating subjects recorded passively.

The submitted algorithm ranked among the top eight most accurate algorithms out of 16 vendors and later contributed to improvements included in VeriLook SDK and MegaMatcher SDK.

In 2018, Neurotechnology ranked sixth of 40 participants in a 12-million-subject identification test and third of 40 when probe images were captured up to 18 years after enrollment.

Neurotechnology algorithms were included in NIST testing for paperless travel and immigration scenarios. Submissions made in 2019 and 2021 were evaluated in simulations involving biometric boarding and airport security galleries.

These tests examined one-to-many identification performance when travelers were compared against galleries created from passport, visa or previous travel images. The evaluation reflects an operational scenario in which face recognition can support passenger processing without repeatedly presenting a physical travel document.

Neurotechnology also participated in the Face Recognition Vendor Test (FRVT), NIST's earlier face recognition evaluation program.

Its algorithms were evaluated in both one-to-one verification and one-to-many identification. In results published on January 13, 2022, Neurotechnology submissions ranked among the leading algorithms in border-control, kiosk, frontal mugshot and profile mugshot scenarios.

In July 2023, FRVT was split and rebranded into Face Recognition Technology Evaluation (FRTE) and Face Analysis Technology Evaluation (FATE), and Neurotechnology continued participating in these newer evaluation tracks.

Active evaluation. In the Face Recognition Technology Evaluation (FRTE), Neurotechnology has participated in both core face recognition and specialized face-mask performance testing.

In FRTE 1:1 Verification results published on November 27, 2023, the company's algorithm ranked among the top results in border-control-related scenarios.

The evaluation compares two facial images to determine whether they belong to the same person. Neurotechnology achieved leading accuracy in supervised and unsupervised scenarios involving visa, border and kiosk images.

Active evaluation. In FRTE 1:N Identification, Neurotechnology showed strong performance in both civil identification and investigation-style scenarios, including frontal and profile mugshot matching and border-control comparisons.

In this evaluation, a probe image is searched against a larger facial image gallery. The tested scenarios include frontal-to-frontal and profile-to-frontal mugshot matching, as well as border-control comparisons involving images captured years apart.

Active evaluation. In FRTE Face Mask Effects results published on December 20, 2023, the algorithm achieved top results for recognizing people wearing face masks or other lower-face occlusions.

The evaluation measures the accuracy cost of lower-face occlusion, which became an operational requirement for many deployments.

Active evaluation. Neurotechnology has also participated in specialized face analysis evaluations under FATE. In FATE Quality, the company's facial image quality assessment algorithm was tested on challenging photographic conditions.

Quality estimation is important because the accuracy of enrollment and subsequent matching depends strongly on factors such as pose, focus, lighting, resolution and facial visibility.

Active evaluation. In FATE MORPH, Neurotechnology participated in the evaluation of face morphing attack detection, an area relevant to document issuance and identity fraud prevention.

The evaluation tests algorithms designed to detect facial images created by combining the features of two or more people. A manipulated photograph of this kind could otherwise be used by more than one person.

Neurotechnology's face presentation-attack detection technology received iBeta ISO/IEC 30107-3 Level 2 certification in December 2025. This independent certification is separate from NIST FATE Presentation Attack Detection and should not be presented as participation in that NIST evaluation.

First place. In 2026, Neurotechnology participated in the Competition on Adapting Foundation Models for Face Recognition Using Synthetic Training Data (AFMFR), held at the 2026 International Joint Conference on Biometrics (IJCB 2026). The submission was entered as DMSTI-Neurotechnology, reflecting a research collaboration between Neurotechnology and DMSTI, the Institute of Data Science and Digital Technologies at Vilnius University.

DMSTI-Neurotechnology ranked first in the Full Data Track, which evaluated large-scale adaptation of the CLIP ViT-L/14 foundation model using approximately 3 million synthetic face images representing 35,000 identities. The submission used full fine-tuning with a Sub-Center ArcFace classification head and achieved the highest average accuracy across the small-scale benchmarks at 95.51%, while leading on IJB-B and IJB-C across all reported false acceptance rate thresholds.

At the strict IJB-C operating point, the submission recorded an 87.42% true acceptance rate at a false acceptance rate of 10⁻⁵, compared with 76.71% for the FRoundation baseline. It also led the RFW evaluation with the highest average accuracy at 91.70% and the lowest standard deviation at 2.33, giving it the strongest results for both accuracy and fairness in the Full Data Track.

Neurotechnology's IREX participation dates back to 2009, when its VeriEye 2.1 iris matching algorithm ranked among the top three for accuracy in NIST IREX I. It was 25–75 times faster than competing algorithms, with templates 3.5–7.5 times smaller.

This established the starting point for Neurotechnology's independently evaluated iris recognition development.

Neurotechnology continued its participation in IREX III, which focused on the performance and interoperability of iris recognition algorithms and compact iris templates.

In 2012, NIST evaluated more than 86 algorithms and variations from nine companies and two universities. Neurotechnology's iris matcher was the second-fastest submission and provided three times higher recognition accuracy than the only faster contender.

The company's iris technology was subsequently evaluated in IREX IV, which examined iris recognition performance using images collected in operational environments.

NIST IREX IV evaluated 66 iris-recognition prototypes from 12 commercial and academic organizations in a large-scale identification task. In the Class P test using a gallery of 1.6 million single-irises, Neurotechnology's submission was the only one to return results in less than 0.5 seconds without reducing recognition accuracy. Across all tests, its submissions ranked among the top four participants for accuracy.

Neurotechnology also participated in IREX IX, extending its evaluation history to more recent iris recognition algorithms and datasets.

NIST evaluated 46 iris-recognition algorithms from 13 commercial and research organizations in IREX IX. Neurotechnology's algorithms ranked second for accuracy, while the accelerated version was nearly 50 times faster than any other matcher in the evaluation.

In October 2019, Neurotechnology joined NIST IREX 10, an ongoing large-scale iris identification evaluation.

IREX 10 accepts submissions continuously and measures one-to-many iris identification at scale.

First place. In 2023 Neurotechnology's iris recognition algorithm has been judged by NIST as the most accurate among the participants in the Rank 1 category.

The submitted algorithm outperformed other contenders in both single-eye and two-eye assessments. Also, it showed top results for most performance metrics.

First place, active evaluation. In May 2026, the company's latest iris recognition algorithm ranked first in all IREX 10 evaluation categories, covering both single-eye and two-eye assessments. The algorithm achieved top results for automated identification and Rank 1 investigation-style accuracy.

The submission achieved the lowest Rank 1 miss rate in both two-eye and single-eye testing, as well as the lowest false negative identification rate at a false positive identification rate of 0.01.

This result also reflects the direction of Neurotechnology's algorithm development: improving not only recognition accuracy, but also matching speed and suitability for large-scale one-to-many identification.

IREX 10 remains an active evaluation.

Neurotechnology participated in the NIST 2024 Speaker Recognition Evaluation, part of an evaluation series conducted by NIST since 1996. SRE24 evaluated automated person detection using conversational telephone speech and audio extracted from video.

The evaluation included audio-only, visual-only and audio-visual tracks, as well as cross-source and cross-language comparisons. It also introduced variable enrollment duration, shorter test segments and recordings containing multiple speakers.

NIST publicly names participating teams but does not link team names to scores, and its evaluation plan restricts promotional claims about standing.

Confirm internally which tracks Neurotechnology entered and which participation details may be published; do not make ranking or endorsement claims.