Andrei Telenkov: ‘India key market for facial recognition tech; hope to grow smart city use cases’

Russian facial recognition know-how supplier NtechLab, which not too long ago introduced a contract with Indian Railways to implement its know-how at 30 stations throughout Gujarat and Maharashtra, sees India as one of many precedence markets for video surveillance and facial recognition know-how. NtechLab’s CEO Andrei Telenkov, in an interview with Pranav Mukul, talks concerning the increasing use instances and the way India is leveraging the know-how. Edited excerpts:
How do you see India as a market?
We undoubtedly see India as one of many key markets, as a part of our worldwide growth. We are already current in India, and have a number of use instances of our product, however the one we made public was the one with Indian Railways. In phrases of measurement, inhabitants and the way nicely we are able to cooperate with companions right here, India is thought internationally for excellent experience in engineering and IT. We are hopeful that our portfolio of use instances will develop — by way of growth with present shoppers — and hopefully we are able to enter completely different segments like good metropolis, retail, and banking. India is certainly a high precedence for us, and this yr we now have employed devoted individuals who will work completely with India.
What are the varied use instances doable in India?
We see an enormous variety of situations and use instances that may be arrange. In Russia, we now have a Moscow Safe metropolis, which we imagine is the most important undertaking worldwide should you don’t rely China. Moscow has 1,70,000 cameras and we don’t do solely facial recognition however completely different type of video analytics. Of course it’s about faces, that’s how we began, however now we now have expanded into completely different applied sciences. We can detect folks by silhouettes, we are able to detect and recognise automobiles even with no quantity plates. We can implement good options on video cameras. We can also recognise weapons in police instances, if somebody is utilizing a weapon in entrance of the digital camera. The variety of instances is rising, rising and rising and we hope that we’ll implement these in different markets as nicely — India being a precedence market. Of course, regulation enforcement is a key situation however there are numerous different instances for instance, cities use video analytics to analyse the circulation of individuals — how many individuals are on the road, what number of are ready for a bus, and so on.
Are there any India-specific challenges that you just confronted?
In phrases of technical challenges, there was a really fascinating problem for our R&D workforce. India was a really arduous case by way of variety of faces. The consumer was Indian Railways, which is arguably the world’s busiest railways. In some geographies, if you do a undertaking of this sort, the typical video may have 10 faces, however in India, in a single body, the typical variety of faces was 50. That was an enormous technical problem to detect and recognise all these faces.
Is it doable for shoppers to make use of your know-how past the acknowledged function?
Absolutely, it is vitally doable. We usually come throughout this case. Many of our shoppers in Russia, they begin with a blacklist — which are the folks they don’t wish to see of their institution. These individuals are suspected to be criminals by regulation enforcement, our shoppers use our know-how to get alerts about these folks in the event that they enter the premises. After step one, they are going to proceed to make use of our system for various situations, different use instances. As an organization, we’re very open to what the consumer places in entrance of us and we test whether or not their drawback will be solved with video analytics.
Can a consumer, for instance Indian Railways, increase its reference database that it makes use of for facial recognition?
This relies on the construction of the contract. Sometimes we’re versatile, the place we may give limitless licence, which will likely be limitless by way of variety of faces it may recognise. But it varies from contract to contract. Technically, it is vitally doable. We can add to and undergo databases very quick. For instance, certainly one of our shoppers, which is an web firm, has a database of 1.5 billion faces and we are able to search a face by way of this big database in lower than 0.3 seconds.
Does an organization like NtechLab construct safeguards so the shoppers don’t use it to violate privateness of residents?
It is de facto as much as the consumer how they use it. From our expertise, nearly all of the shoppers are established establishments, which have their compliance techniques in place. From what we all know, the shoppers are very cautious as a result of that is very delicate data. We don’t get the entry to the database, we simply present the know-how and the know-how however we can’t touch upon every consumer as a result of we’re know-how suppliers. For instance, Indian Railways is a good consumer and this has extraordinarily low likelihood of taking place. We don’t understand how the system is used exactly, we offer the know-how and it’s as much as the consumer to compile a database and test on this database.