“Partnering with NtechLab introduced flawless facial recognition technology into our software. We felt supported by experts every step of the way. Each day, retail chains in Yekaterinburg, Moscow, and other cities in Russia prevent and solve dozens of crimes.

On average, a retail chain’s losses are reduced by 2−3% of annual revenue, while the cost of acquiring the software sees a full return on investment in 2−4 months. Integrator costs for acquiring a FindFace license are usually recouped after the first deal.”

Mikhail Korablyov,

CEO of BIT Ltd.

Our partner

The Urals-based company BIT has developed software and offered consulting services for over 5 years, specializing in security issues. In their efforts to create their own facial recognition software product for retailers, BIT employees spent a long time testing various algorithms before selecting a partner that would oversee facial biometrics. BIT has collaborated with NtechLab since 2016.

Objectives

  • Create surveillance software to prevent shoplifting in retail stores and chains. Process video feeds from multiple cameras, detect shoplifters quickly and accurately, and keep the recognition output high even under compromised lighting conditions.
  • Implement the software on existing general-purpose hardware.
  • Ensure minimum restructuring of the customer’s current business operations.
  • Obtain acceptable results with low-res cameras and from compromised viewpoints.

Solution

Using FindFace technology, BIT’s engineers created the Stop Shoplifter solution that supports facial recognition in live and recorded video combined with separate data storage in retail chains. The solution also supports checking against a centralized offender database and passing on forensic videos to the police forces.

The result

The NtechLab’s FindFace powered Stop Shoplifter solution checks an offender against a database containing ten thousand entries in no more than 2 seconds.

The unique FindFace algorithm boosts recognition accuracy to nearly 100% while processing thousands of camera channels simultaneously. This makes the software solution both efficient and highly competitive.

As a result, the technology has seen an increased demand in retail, where theft is especially high. Large grocery and perfume retailers in Yekaterinburg were the first clients to embrace the solution. The results are impressive: theft attempts are prevented every day while over 15 repeat offenders are handed over to the police each month. Theft cases that previously went unsolved due to not being able to find or identify the offenders are now being successfully investigated. Shortly after launching the program, customers decided to scale the project. They installed the software in the Moscow region, in Kazan, Novosibirsk, Saint Petersburg, and other large cities.

This solution allowed retailers to reduce losses by millions of roubles while also preventing reputational losses.