Parts | Days | Selection | Search | Updates | Downloads | Help

T: Fachverband Teilchenphysik

T 21: Data analysis, Information technology I

T 21.4: Talk

Monday, March 15, 2021, 16:45–17:00, Tu

Improved energy resolution via super-resolution — Johannes Erdmann, Florian Mentzel, Olaf Nackenhorst, and •Aaron van der Graaf — TU Dortmund, Experimentelle Physik IV

In high energy particle physics, detectors with a good energy resolution are essential for the precision of measurements. One possibility to improve the energy resolution are detector upgrades. Another approach is to artificially enhance the energy resolution by using super-resolution (SR) algorithms. SR algorithms learn to upscale low resolution data to high resolution data. The SR algorithms that are used in this work are based on generative adversarial networks (GANs). By training GANs with simulation-based high resolution and low resolution data, they have been shown to learn the complex correlations between low and high resolution data. After the training, GANs can then upscale the low resolution data. In this presentation, preliminary results are presented.