Skip to main content Skip to main navigation


Generation of synthetic training data for AI models

Synthetic data can be generated in large quantities faster and cheaper than their counterparts and are used in many applications in AI development. As training material for AI models, they provide the data basis for new experimental use cases. In the Agri-Gaia project, the research groups of the Plan-based Robot Control and Marine Perception research departments at DFKI have developed a new approach for detecting plastic waste in the environment, especially in agricultural settings.

In this use case, realistic computer graphics are used to generate artificial images whose contents, such as the positions of individual objects in the image, are known and can therefore be easily used for machine learning. At Hannover Messe, the project team from DFKI's Lower Saxony site will explain the underlying problem and demonstrate its method and possible applications. In Agri-Gaia, DFKI and partners from academia and industry are exploring open AI technologies for the agri-food industry within an AI ecosystem based on GAIA-X, the European sovereign data infrastructure.

Agri-Gaia is funded with 11.75 million euros by the German Federal Ministry of Economics and Climate Action over a period of three years.

More information

Visit us at the:
MWK Lower Saxony Joint Booth
Hall 2, Stand A40

Project Website