Prague, July 2 (CTK) – Scientists from Czech Technical University in Prague (CVUT) have created a system able to recognise mushrooms and plants species from photos, the CVUT said in a press release on Monday.
The system succeeded at the international competition ExpertLifeCLEF 2018.
Milan Sulc and Jiri Matas, from the Department of Cybernetics of the CVUT Faculty of Electrical Engineering, created an automatic system able to distinguish the species of mushrooms or plants from their shape and colour in photographs.
Lukas Picek, from the Department of Cybernetics of the Faculty of Applied Sciences of University of West Bohemia (ZUC), also took part in the project.
The system is based on deep neural networks and is able to learn and to improve.
At the international competition, the individual systems were competing in the set of 10,000 plants they had to evaluate.
The system of the scientists from CVUT and ZUC gained the first place as it was more accurate than most experts.
Now the scientists from CVUT and CZU are considering two trajectories of their further work.
Either they will develop their own application to recognise plants and mushrooms or they will start offering their know-how to other development workers and providers of online services, CVUT said in its press release.