DPG Phi
Verhandlungen
Verhandlungen
DPG

Regensburg 2019 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

MM: Fachverband Metall- und Materialphysik

MM 37: Topical session (Symposium MM): Big Data Analytics in Materials Science

MM 37.7: Topical Talk

Donnerstag, 4. April 2019, 17:15–17:45, H43

Microstructure is the know-it-all - classification approaches with data mining and deep learning methods — •Frank Mücklich1,2 and Dominik Britz2,11Universität des Saarlandes, Saarbrücken, Germany — 2Material Engineering Center Saarland, Saarbrücken, Germany

The microstructure can be regarded as the *multi-scale archive* from which we can *read* the quantitative information about the microstructure formation processes and the prediction of the final material properties on each relevant scale. Recent advances in 3D tomography methods on the micro, nano and atomic scale allow to study the differences of microstructures with higher morphological and topological complexity. Classification strategies using Support Vector Machine and Deep Learning will be discussed using morphological and substructure parameters. Images are processed directly in the workflow after an adapted contrasting. The result might be simultaneously segmented and classified.

S. M. Azimi, D. Britz, M. Engstler, M. Fritz, and F. Mücklich, Advanced Steel Microstructural Classification by Deep Learning Methods, Scientific Reports (Nature) 8 (2018) 2128.

100% | Mobil-Ansicht | English Version | Kontakt/Impressum/Datenschutz
DPG-Physik > DPG-Verhandlungen > 2019 > Regensburg