PJSC Sberbank : Materials from Machines Can See 2021 published

VisionLabs has held the fifth Machines Can See international computer vision and machine learning summit. This year's summit was held in a hybrid format - both online and offline - for the first time. The business and research speakers' presentations are now available on VisionLabs' official YouTube channel, while the research segment is also available on the machinescansee.com website.

The main theme of this year's summit is human-centric technology, which focuses on analytics based on silhouettes, gestures, poses, gaze direction, and other attributes without collecting and analyzing personal data. As part of Machines Can See's scientific program, researchers from leading world universities and industrial AI labs discussed recent studies in this field.

· Dima Damen, University of Bristol

· Efstratios Gavves, University of Amsterdam

· Bernard S. Ghanem, King Abdullah University of Science and Technology

· Ira Kemelmacher-Shlizerman, University of Washington; UW Reality Lab; Google

· Kris Kitani, Carnegie Mellon University

The business segment of the summit focused on the role of computer vision in business' digital transformation. Experts from Sber, MTS AI, Huawei, Alfa Bank, and other major companies discussed the effectiveness of the technology using successful cases studies as examples.

An international data analysis competition was held from June 7 to July 5 in the framework of the summit. The participants had to create an algorithm for recognizing one of six gestures - Stop, Victory, Mute, Ok, Like, Dislike.

Fifty-seven teams competed, demonstrating a high level of sophistication in their solutions: over half the teams surpassed the basic algorithm, and the winners improved its results by a factor of 3.5. The prizewinners' algorithms are capable of recognizing gestures in real time, which is particularly important when using technology to control smart devices. The prize fund for the competition was RUB 750,000.

According to AI Index Report, one-third of all scientific publications on AI in 2020 focused on computer vision. Since 2015, the number of articles on this topic written per year has increased sixfold. Computer vision is one of the most dynamically growing areas of AI. Over five years, Machines Can See has become one of the most significant international events in the field. Over 3,000 people from 91 countries took part in this year's summit. The average viewing time surpassed two hours. This speaks, on the one hand, to the interesting and high-quality speech content, and, on the other hand, to the fact that everyone made a conscious choice to attend. The summit brought together true leaders in the industry - top Russian and world experts in computer vision and representatives of major companies that are implementing this technology.

Alexander Vedyakhin

First Deputy Chairman of the Executive Board, Sberbank

VisionLabs is a world leader in creating products powered by computer vision and machine learning. The company counts financial and transport companies, retailers, and international tech corporations among its clients. Its main products are the LUNA computer vision and video analytics platform and a device line based on its own technology (Luna Ace, Luna Pos, Luna Thermo, and others).

VisionLabs' algorithms are consistently ranked highly in regular independent tests carried out by the US National Institute of Standards and Technology (NIST).

Sberbank and venture fund Sistema_VC are two of the company's investors. VisionLabs is a resident of the Skolkovo Foundation.