Russian Engineers Receive 1.5 Million Rubles as Winners of Ice Vision Hackathon

On July 16, the Skolkovo Institute of Science and Technology (Skoltech) hosted the award ceremony for Ice Vision, an international hackathon for driverless vehicle software developers. The first place went to young professionals from NtechLab, followed by student teams from the Moscow State University, Higher School of Economics, Siberian State State University, Skoltech, and Novgorod State University. Ice Vision is part of the Up Great technology competition series held under the auspices of the National Technology Initiative.



With 26 teams representing six countries, the 48-hour long hackathon was held simultaneously at the Moscow Institute of Steel and Alloys and Skoltech. 

The contestants were required to design AI-based solutions to help driverless vehicles better analyze road conditions in winter weather. To do so, they were offered a “winter” dataset (with over 600 thousand images showing roads under harsh weather conditions) and lidar data that enables the creation of real-time 3D maps of near-vehicle environments. 

The best solution was provided by the team Azat and Artem (representatives of NtechLab) who were awarded with the main prize of 1.5 million rubles. The joint MSU-HSE team PsinaDriveNow was second and received 300 thousand rubles, while Vizorlabs from the Siberian State University finished third and received 200 thousand rubles.

“In our work, we used convolutional neural networks. We believe that this was the right decision. By the way, after we had finished, we discussed our solution with the team that took second place, and it turned out that they moved in a similar direction. They even showed us a few gimmicks that we could use to improve our final result,” said Azat.

The hackathon's partners provided additional cash prizes for the contestants. The top 10 teams received 100 thousand rubles per team from StarLine, while the teams that placed 11th through 15th received individual NVIDIA DLI vouchers. 

“Driverless transportation presents great opportunities and challenges to the developers of artificial intelligence systems around the world. By running a hackathon focused on computer vision solutions  for autonomous vehicles we move a step closer to the creation of Level 5 automated driving systems. I am sure that the solutions proposed by the young engineers participating in Ice Vision will find practical application by our business partners,” commented Alevtina Chernikova, provost at the Moscow Institute of Steel and Alloys.

At the hackathon, students from Russian universities competed with their peers from the United States (Arizona State University and the University of Wisconsin-Stout), China (the University of Science and Technology Beijing and the Harbin Institute of Technology),  Spain (the Technical University of Catalonia), France (the University of Paris-Saclay), and South Korea (the University of Ulsan).

17 july 2019