NEC NeoFace
NEC Corporation · Japan
Confidence 3/5
Technical capabilities
NEC NeoFace is a family of face-recognition products that have consistently ranked among the most accurate algorithms in successive NIST Face Recognition Vendor Test (FRVT) evaluations since 2010. The product family includes NeoFace Watch (real-time camera-feed recognition), NeoFace Reveal (forensic still-image search), NeoFace Express (border-control), and NeoFace Reveal Mobile.
NEC publicly claims accuracy of more than 99 percent under controlled conditions. Real-world deployments, including the London Metropolitan Police Live Facial Recognition (LFR) trials between 2016 and 2019, have reported substantially higher false-positive rates in operational conditions, with independent academic evaluation by the University of Essex finding 81 percent false-positive rates across eight LFR trial deployments.
Documented use
- London Metropolitan Police: NeoFace Watch deployed in Live Facial Recognition trials at Notting Hill Carnival, Remembrance Sunday and Stratford between 2016 and 2019, then operationalized from January 2020. Big Brother Watch and Liberty challenged the deployments in UK courts.
- South Wales Police: NeoFace deployed with similar Bridges v. South Wales Police judicial review (UK Court of Appeal, August 2020) finding partial unlawfulness in deployment criteria.
- Argentina: deployed in Buenos Aires by city government from 2019, suspended in 2022 after court ruling on misidentification of more than 130 individuals.
- India: deployed by Delhi Police, Hyderabad and other state forces.
- US Customs and Border Protection: airport biometric exit program (TVS / Traveler Verification Service).
- Australia, Brazil, Japan: multiple government and airport deployments.
Customer states
Government and airport-operator customers in at least 70 countries per NEC investor relations. The vendor states adherence to Japanese export-control regulations and to in-house human-rights principles published in 2021. Independent evaluation of those commitments is limited.
Legal and sanctions status
- Not on the US Department of Commerce Entity List.
- Not designated by US OFAC.
- Bridges v. Chief Constable of South Wales Police (UK Court of Appeal, August 11, 2020): partial unlawfulness in operational selection criteria.
- Argentine federal court order (March 2022): suspended Buenos Aires deployment.
- Subject to Japan METI export-control regime.
Technical countermeasures
- Face occlusion: surgical masks, scarves and sunglasses reduce recognition accuracy substantially. Pandemic-era mask usage degraded NeoFace performance to NIST-measured loss of accuracy.
- Oblique angles: side-view or downward gaze degrades probe-quality and reduces match confidence.
- Adversarial accessories: research projects (CV Dazzle, anti-recognition makeup, Reflectacles glasses) have demonstrated partial effectiveness against deployment-grade systems.
- Reduce live-camera exposure: in jurisdictions with LFR deployment, situational awareness of camera positions allows route planning.
- Legal challenge: subject-access requests, FOIA equivalents, and judicial-review actions have proven effective at limiting LFR deployments in the UK and Argentina.
Sources
- NIST FRVT, Ongoing benchmark of face-recognition algorithms
- Big Brother Watch, The State of Facial Recognition in the UK
- University of Essex, Independent report on Met Police LFR trials (July 2019)
- UK Court of Appeal, Bridges v. Chief Constable of South Wales Police (August 11, 2020)
- Liberty, Resist Facial Recognition campaign
- Reuters, Argentina court suspends Buenos Aires facial recognition (March 2022)
Update log
May 1, 2026: Entry created. Sourced from NIST FRVT independent evaluation, Big Brother Watch and Liberty campaign documentation, University of Essex academic report on Met Police LFR trials, UK Court of Appeal judgment in Bridges, and Reuters coverage of the Argentina judicial suspension.
There’s no perfect setup. Anyone selling you perfect is selling fear. The goal is simple: make yourself a harder target than the person next to you.
