D1.3 Ensemble Learning for Computational Optical Form Measurement
- Event
- SMSI 2021
2021-05-03 - 2021-05-06
digital - Band
- SMSI 2021 - System of Units and Metreological Infrastructure
- Chapter
- D1 Future Topics in Metrology (Special Session)
- Author(s)
- L. Hoffmann, I. Fortmeier, C. Elster - Physikalisch-Technische Bundesanstalt, Braunschweig/Berlin (Germany)
- Pages
- 318 - 319
- DOI
- 10.5162/SMSI2021/D1.3
- ISBN
- 978-3-9819376-4-0
- Price
- free
Abstract
Deep learning has become a powerful tool of data analysis with applications in such different areas as medical imaging, language processing or autonomous driving. Recently, deep learning techniques have also been applied to an inverse problem in optical form measurement. In a proof-of-principle study it was shown that an accurate solution of the inverse problem can be achieved by a deep neural network that is trained on a large data base. This work augments the developed method with a quantification of its uncertainty by considering an ensemble of networks. The approach is tested using virtual experiments with known ground truth.