FOR LARGE-SCALE IMPLEMENTATIONS
Top-performing algorithm at identification task on 12 mln database according to National Institute of Standards and Technology
MULTIMILLION DATABASE REAL-TIME SEARCH IS REAL NOW
The algorithm applied to NIST 1:N 2021 is implemented in a country with 30 mln population to perform a throughout search.
database is deployed for real-life scenarios
Search hundreds of target objects simultaneously without additional hardware
FASTEST IN THE MARKET
Reconciliation of chip architecture in optimization problem within ML tasks results in the best performance without accuracy trade-off.
Template generation time per 1 CPU core.
Search duration for 12M database per 1 CPU core.
COUNTRY-SCALE LAW VIOLATORS
Study a case of the offence even though the video quality is poor. Track people movement history throughout different cameras, date and location. Get real-time mobile alerts once the target object detected.
Identify your customer once she enteres your door. Personalise greeting, recall shoes size, suggest bevarage based on menu preferences and many more.
In-store personilised service and conversation with every customer is a key to superior customer experience.
DEVICELESS ACCESS TO PUBLIC TRANSPORTATION
No turnstile, ticket conductor or card-reader in the cabin. Make your customer inter-city trip as convenient as it could be, increasing throughput, transparence and security of the municipal transportation system.
Face payments is the future, but the world still lacks NFC terminals, a previous generation hardware. So why catch-up with the rolling out of already outdated hardware? Tablet with a cheap web-camera is the payment hardware of the future.
We are a team of computer vision developers
with 20 years experience in the field.
Focusing on country-scale applications, we developed the unique proprietary AI framework to make face recognition work for 20+ mln. databes.