The company NtechLab, a world leader in accuracy and speed of face recognition, silhouettes, and action on video, reports on the victory of the team in international competition in the field of machine vision Ice Vision. This is stated in the press release received by the editors”.ru”.
Software solutions, valued in the competition, will allow unmanned vehicles to effectively and accurately detect a road situation in the Russian winter. In the final of the competition attracted 26 teams from around the world.
Their developments in the field of recognition of the traffic situation, the finalists were tested on a unique “winter” dataset, which was developed by the specialists of the “Skoltech” and contains more than 600 000 images taken under difficult weather conditions, as well as data, allowing to create real time 3D map of the space surrounding the vehicle. The struggle for the main prize took place between the Russian developers and specially arrived to Moscow from USA, China, South Korea and Spain. The team, representing the company NtechLab, won first place with a prize Fund of 1.6 million rubles.
“In our work we used a convolutional neural network. Believe that it was the right decision. By the way, later, when we moved away from the computers, we discussed our decision with the team, which took second place, and it turned out that they acted approximately on the same principle. They even showed us a few things that could improve our final result”, — told the team members.
The development of unmanned vehicles is one of the most pressing challenges facing developers of systems based on neural networks throughout the world. Holding a hackathon aimed at finding effective solutions in the area of improving computer vision for Autonomous vehicles, allows to approach the creation of unmanned systems 5th level, the introduction of which does not require human intervention in the process of driving.
“NtechLab, has established itself as the leader in accuracy and speed of face recognition, is working to expand the range of applications of neural networks. In 2017, developed by our team algorithms have started to recognize the emotions of the people in 2018 we have been among the world leaders in the recognition of silhouettes of people, in 2019 has brought leadership is by definition action in the video. Technological superiority in the recognition of the road environment opens new perspectives in enriching the Analytics of “smart” and “safe cities” unique data”, — commented the founder and head of the laboratory of neural networks NtechLab Artem Kukharenko.
Competition Ice Vision — part of a series of technological competitions Up Great, held for the purpose of implementation of the National technology initiative. Co-organizer of the contest has been one of the leading Russian technical universities nust “MISIS” strategic partner – research-and-production Association “STARLINE”, technological partner — Sktech. The event was held in the format of a 48-hour hackathon on the grounds of the “House-commune” nust “MISIS” and SKOLKOVO Institute of science and technology (“Skoltech”).
NtechLab is the world leader in accuracy and speed of face recognition, silhouettes, and actions on video. Algorithms NtechLab repeatedly proven its technical superiority in representative international competitions. In 2017, the development NtechLab took first place in the competition of the American Agency for advanced studies in exploration in the categories of “most accurate” and “fastest” algorithm. In 2018 NtechLab among the three winners of the contest WIDER Pedestrian Challenge on the detection of pedestrians based on their silhouettes. The company’s products are in demand in the areas of public and corporate security, retail trade, financial sector, entertainment industries and hospitality. The algorithm is being successfully piloted in 10 large cities of Russia and CIS countries, as well as municipal surveillance systems of India, South America, the Middle East and Southeast Asia.
Video, photo All from Russia.