CYBATHLON is a competition organized by ETH Zurich, one of the most prestigious technical universities in Europe. The competition, held every 4 years, is considered as a Formula 1 race for devices designed to help people with disabilities and improve their quality of life.
In the competition, teams consisting of technology developers from universities, companies or NGOs and a person with disabilities tackle various everyday tasks using their latest assistive technologies.
Since its launch in 2013, more than 100 teams from over 30 countries have participated in CYBATHLON competitions, which are taking on new forms and include an increasing number of disciplines.
One of the new categories launched this year is a race with smart visual assistance technologies for people with severe visual impairment.
The competition consisted of two tasks that are common in everyday life. Firstly, the pilot had to pass through an obstacle course testing spatial orientation without touching any objects exposed as obstacles. The second task was to set a table by placing a plate of soup between the cutlery without touching it and without spilling a drop of soup.
In this new category, the "EyeRider" team from the Bionic Vision Centre at Pázmány ITK also competed with an improved version of their LetSee mobile app. The new feature tells the names and positions of the objects in the camera image, using an engine based on a deep neural network.
The EyeRider team's blind pilot, Péter Sulyok, has been involved in testing the app and developing the user interface and interactions throughout the preparation phase. His outstanding situational awareness and intelligence were key to the team's success.
The team was led by Kristóf Karacs, Head of the Bionic Vision Centre, and Kristóf Iván, CYBATHLON Coordinator at ITK. The lion's share of the algorithm development was done by Anna Gelencsér-Horváth, PhD candidate, who also coordinated the work of the students who joined the project in February this year (Subham Shome, Yu Hyeon, Ákos Madaras, Zsolt Vastag Botond). Viktor Karacs, project coordinator, was responsible for the design of the training ground and the training of the pilot. Gábor Quintz, who has been the main developer of LetSee for many years, provided significant support in mobile programming.
The work was also supported by Pázmány University and the ELTE Faculty of Informatics, as well as the professional cooperation with the NIPG research group of the ELTE Department of Artificial Intelligence, with the contribution of András Lőrincz and Viktor Varga.
The teams were given two opportunities to solve the tasks, with EyeRider's Peter completing the course flawlessly in the second round to win the competition.
Congratulations to the whole team for a great performance!
