Neuronal extraction of information, structures and symmetries in images

Image analysis with the help of artificial intelligence and machine learning

 

Proof of concept: Deep neural networks can accurately reconstruct plasma parameters (here a proxy for the rotational transform at the edge) from heat load images which will aid machine control.

Max Planck Institute for Plasma Physics is participating in the NEISS project (Neuronal Extraction of Information, Structures and Symmetries in Images) in a research alliance with the University of Rostock. The focus of this interdisciplinary project, which is part of the excellence research program of the state of Mecklenburg-Vorpommern, is the analysis of images using artificial intelligence and machine learning.

One of the things being worked on in the project's plasma physics work package is a control system for Wendelstein 7-X involving measurements of not just one but several diagnostic systems. This includes, for example, images from infrared or X-ray cameras or data supplied by spectroscopic or magnetic diagnostics. Control signals are to be obtained from these in real time and fed back to the machine. Such a real-time control system should guarantee optimal and reliable operation of the machine.

 

The NEISS project is funded by the Excellence Research Program of the state of Mecklenburg-Vorpommern with contributions from the European Social Fund (ESF) of the European Union. 

Go to Editor View