Step 1
1. We find the face in the frame
The neural network detects the face in a photo or video, even when several people are in frame, and aligns it by landmark points.
A face recognition system for business
The NeuroVision face recognition service compares a face with a photo or finds it in your database in 0.1 seconds. The best result among Russian vendors in the international NIST FRVT benchmark.
Top-1 face verification algorithm in Russia, top-24 worldwide — NIST FRVT 03/2023*
Data is stored in Russia Russian Software Registry No. 19972 Compliance with 152-FZ
* The algorithm is in the top-25 of NIST FRVT (March 2023) — the best result among participants from Russia
What NeuroVision face recognition can do
A single biometric engine for comparing, searching and checking a face — it fits into onboarding, app login, an entrance gate or a checkout.
Comparing a face with a photo (1:1 verification)
Is it the same person? A selfie is compared with the passport photo or with a reference image captured earlier.
Searching a face database (1:N identification)
Who is this? The face is searched in your own database: among employees, customers or in a stop list.
Biometric authentication
Logging into an app, confirming a payment or a change of details by face instead of an SMS code or a password. It works with an ordinary front camera.
Face attributes: age, sex, emotions
Age estimation from the face for 18+ sales and age-restricted services; sex and emotions for footfall analytics.
Liveness detection
It tells a live person from a photo, a video, a mask and a deepfake. More about liveness detection
How face recognition technology works
Four steps from the frame to the decision. More on how face recognition works
Step 2
2. We check the image quality and liveness
Sharpness, lighting, head angle; and liveness where needed, to filter out screen photos and deepfakes.
Step 3
3. We build a biometric template
The face is turned into a numeric vector — a biometric template that the comparison then works with.
Step 4
4. We compare and return the result
The template is compared with a reference (1:1) or with a database (1:N) in 0.1 sec. The API returns a similarity score and a decision.
Face recognition accuracy according to NIST FRVT
- 01
What NIST FRVT is
NIST is the International Institute of Standards and Technology. In the FRVT benchmark (Face Recognition Vendor Test, now FRTE) algorithms from vendors worldwide are tested on the same closed photo databases. It is the best known independent ranking of face recognition accuracy. The NIST FRTE ranking
- 02
Works in low light, with a mask and with make-up
The 99.74% accuracy holds when a person has changed their hairstyle, grown a beard, put on make-up or a medical mask, or is standing in a dimly lit room. FAR, FRR and the NIST results
NeuroVision results
NIST FRVT benchmark, March 2023
- № 1 among Russian vendors
- top-24 in the global ranking
- 99,74% recognition accuracy
Where a face recognition system is used
One engine — different scenarios. Details of each solution are on its own page.
- Learn more
Remote customer identification (KYC)
A selfie is compared with the passport photo during online sign-up.
- Discuss your scenario
Face-based access control
Passing a turnstile without a pass or a card.
- Discuss your scenario
Biometric time and attendance
Clocking in and out with no «buddy punching».
- Discuss your scenario
Payment by face
Paying at the till or in the canteen without a card or a phone.
- Discuss your scenario
Video analytics
Visitor counting, stop lists and footfall analytics on your own cameras.
- Discuss your scenario
App login and transaction confirmation
Contactless authorisation and payment confirmation instead of SMS.
Face recognition for your industry
The problems biometrics solves in each industry.
-
Banks and fintech
- Remote account opening, confirmation of large transfers, protection against logging into someone else's account
- A 1:N search finds the same person across several loan applications
-
Retail and marketplaces
- Payment by face, age verification at self-checkouts
- Verification of sellers and couriers
-
Office buildings, warehouses and manufacturing
- Entry by face and shift tracking
- Access to hazardous areas only for employees who have been briefed
-
Transport, car sharing and delivery
- Checking that the person driving or taking the order is the registered one, not someone the account was passed to
-
Online services and 18+ platforms
- Age verification and one account per person
- Protection against multi-accounting and bonus abuse
-
Education and online exams
- Confirming a student's identity in an online exam and on entry to a lecture hall
- Protection against exam impersonation
Why a business needs biometric identification
-
It cannot be passed on
A password, a card and an SMS code can be handed over or stolen; a face cannot.
-
Faster for the customer
A glance at the camera instead of typing a code or looking for a pass.
-
Less fraud
Repeat sign-ups, money mules and logins into someone else's account are detected automatically.
-
Cheaper than SMS and passes
There is no need to pay for SMS codes or print access cards.
Biometrics replaces passwords, SMS and passes wherever it matters that the action is performed by this particular person.
Security of biometric data
Work with biometrics meets Russian legal requirements and those of your security team.
-
We store a template, not a photo
A numeric face template is used for comparison. The original images need not be kept.
-
Data in Russia and inside your perimeter
A cloud on servers in Russia, or on-premise deployment where biometrics never leaves your infrastructure.
-
User consent and legal requirements
The SDK shows a consent screen before capture. Biometric workflows are configured to meet the requirements of 152-FZ.
Why NeuroVision
Face recognition is the core of the NeuroVision platform: the same engine works in KYC, anti-fraud and access control.
NeuroVision — an expert in computer vision and AI
We provide KYC verification services trusted by the largest companies in fintech, e-commerce, the crypto industry and transport services.
We develop identification algorithms that automate checks, detect fraud and help meet regulatory requirements.
The best in Russia according to NIST
Top-1 in Russia and top-24 worldwide
A Russian developer
Skolkovo, Software Registry No. 19972
Face recognition API and SDK: live in a single day
A cloud API, a ready-made capture interface or deployment inside your perimeter — face recognition fitted to your architecture.
- Similarity and decision in JSON
- 1:1 and 1:N in one API
- Built-in liveness
- Sandbox and documentation
Terms of connection
A free one-month test, a dedicated manager, 24/7 support, payment only for checks performed.
Licences and certificates
The project «Development of a multimodal system using artificial intelligence "Neuromed" for managing the diagnostic processes of a clinic on the basis of the "Neurovision" computer vision platform» is supported by the Innovation Promotion Foundation within the federal project «Domestic Solutions»
The Neurovision computer vision platform is registered with Rospatent (certificate No. 2023661508) and included in the Russian Software Registry (entry No. 19972 of 13.11.2023)
The right to use the Platform is granted under a licence agreement concluded by Neurovision with its customers
Test face recognition for free
Leave a request — we will give you a month of test access to the API and help you choose a scenario.
Can't put your request into words? Book a call and our team will contact you at a convenient time
Book a call
The company is included in the register of IT companies and the register of personal data operators of the Russian Ministry of Digital Development.
Address: Moscow, Presnenskaya Embankment 12 — Moscow City business centre, Federation Tower. Nearest metro stations: Delovoy Tsentr and Mezhdunarodnaya.
Read also
Materials on face recognition, algorithm accuracy and protection against deepfakes — in our blog.
50 29
KYC, AML and Digital Identification
115-FZ: What Businesses Are Required to Do and How to Meet the Law’s Requirements
107 30
KYC, AML and Digital Identification
Politically Exposed Person (PEP): Who It Is and What 115-FZ Requires
138 16
KYC, AML and Digital Identification
Beneficial Owner (UBO): Who It Is, How to Establish One, and Why It Matters
92 30
KYC, AML and Digital Identification
Due Diligence: What It Is and How to Vet a Counterparty
Frequent questions about face recognition
Answers about the technology, accuracy, spoofing protection, data storage and integration
The technology is called biometric face recognition. Comparing two images of the same person is called verification (face verification); searching for a person in a database is called identification.
1:1 checks that a person is who they claim to be. 1:N works out who they are among all the faces in the database.
NeuroVision's accuracy is 99.74%. In the independent NIST FRVT benchmark the algorithm showed the best result among Russian vendors and entered the global top-24.
A liveness check works alongside recognition: it tells a live person from a photo, a video on a screen, a mask and a deepfake. More on the liveness page.
Yes. The algorithm keeps its accuracy through changes of appearance, make-up, a medical mask and low light.
On servers in Russia, or in your own infrastructure with on-premise deployment.
For online scenarios the front camera of a smartphone or laptop is enough. We will select hardware requirements for access control and tills to suit your site.
You pay per check, and the first month is a free test. On-premise pricing is calculated individually.
