Reliable Data for Autonomous Driving

14.04.2024|12:43 Uhr

Reliable data for reliable artificial intelligence - that is the topic of the new BMBF joint research project RELiABEL. Wuppertal scientists led by IMACM member PD Dr. Matthias Rottmann are working together with the technology start-up QualityMatch GmbH from Heidelberg on new methods that can be used to reliably label image data for autonomous driving.

Reliability is one of the biggest challenges in autonomous driving. The better an AI understands the data of a car's surroundings, the safer its next step will be. This is where the RELiABEL project, funded by the German Federal Ministry of Education and Research, comes in.

The image shows two challenges that researchers are facing in autonomous driving. The images on the right have a low resolution so that the AI cannot clearly determine the image content. In the images on the left, manual annotations are incorrect or missing, which means that the AI is trained incorrectly.

The consortium is now working on methods with which data can be described efficiently and reliably at the same time. On the one hand, data points in the images are labeled multiple times, while on the other hand, manual annotations are saved and reliably supplemented using mathematical methods.