Face recognition system – Newspaper Kommersant No. 41 (7486) of 03/11/2023

Russian developer of facial recognition systems NtechLab is testing a similar system for recognizing wild animals. There is a demand for such solutions in Russian cities where there have been incidents, for example, with the invasion of brown bears. Market participants admit that the system will make it possible to more humanely prevent conflict situations with animals.

The developer of face recognition systems based on artificial intelligence NtechLab (among the co-owners are Rostec and the Ruben Vardanyan Foundation) is testing a wild animal recognition system, the company told Kommersant. NtechLab, together with the Yugansk Nature Reserve, the Malaya Sosva Nature Reserve, and the Quantorium Children’s Technopark in Nefteyugansk, integrated video with bears into the system. Based on this base, the system will distinguish a bear approaching a settlement from any other object and send a notification to the workers of the reserve, and they, in turn, will be able to drive it away to a safe area without eliminating the animal, NtechLab explains.

The first testing of the system is scheduled for April in the Pionersky settlement, located near the city of Yugorsk, and the start of commercial operation is scheduled for 2024. NtechLab also plans to integrate recognition of other animals into the system. According to SPARK-Interfax, in 2021 the company’s revenue amounted to 1 billion rubles. with a net profit of 88 million rubles.

“Now we are seeing more and more requests from the market for animal recognition systems, especially within the framework of intelligent transport infrastructure,” says Dmitry Markov, CEO of the developer of recognition systems VisionLabs. The NtechLab system may be relevant for regions such as the Trans-Baikal and Kamchatka Territories, the Republics of Komi and Khakassia, although last year, even in the Leningrad Region, wild animals were recorded in settlements in winter, recalls Alena Grishkova, a lawyer for KA Pen & Paper.

According to the Quantorium children’s technopark in the city of Nefteyugansk, in 2022, 39 brown bears invaded the settlements of Khanty-Mansi Autonomous Okrug-Yugra, eight of them were killed due to a threat to society. Most predators in 2022 were seen in the settlements of the Nizhnevartovsk region (36%). In 2021, 49 brown bears invaded the settlements of KhMAO-Yugra, 26 of them were killed. In Yakutia, in 2022, cases of brown bears entering settlements became more frequent, after which the local government issued 32 decisions on the extraction of 43 bears that pose a threat to the life and health of citizens.

It is not difficult to train neural networks to recognize animals, notes Irina Ostroukhova, director of work with partners at RecFaces, but the question of comparability of costs and effectiveness remains. “How many cameras do you need to hang to close the entire perimeter of the settlement? At the same time, animals mainly come out at night, which significantly increases the requirements for cameras,” she points out.

Now a similar project is being created in Canada – sensors register a polar bear and report its approach to residential buildings, says Varvara Levitskaya, chief coordinator of projects for the conservation of biodiversity in the Arctic of the World Wildlife Fund (WWF Russia, included in the register of foreign agents by the Ministry of Justice). According to her, it is still difficult to assess the effectiveness of the NtechLab system. “So far, most likely, it is about demonstrating the idea to the market,” says Irina Ostroukhova.

The Ministry of Natural Resources told Kommersant that they supported “the minimization of possible conflict situations between animals and humans,” but the ministry did not hear about the NtechLab project: “Localization of hunting resources that have entered the settlement is within the powers of the police.”

Timofey Kornev