Multi-Task Deep Learning for Surface Metrology
Sensors · 08.12.2025 · 25 (24) · 7471
DOI: 10.3390/s25247471 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 08.12.2025. Verfasst von Dawid Kucharski und 11 weiteren Autor:innen. Beteiligte Einrichtung: Division of Metrology and Measurement Systems, Institute of Mechanical Technology, Faculty of Mechanical Engineering, Poznan University of Technology…. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 31 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministry of Education and Science of Poland.
Abstract (Originalsprache)
A reproducible deep learning framework is presented for surface metrology to predict surface texture parameters together with their reported standard uncertainties. Using a multi-instrument dataset spanning tactile and optical systems, we jointly address measurement system type classification and regression of key surface parameters—arithmetic mean roughness (Ra), mean peak-to-valley roughness (Rz), and total roundness deviation (RONt)—alongside their reported standard uncertainties. Uncertainty is modelled via quantile and heteroscedastic regression heads, with post hoc conformal calibration used to obtain calibrated prediction intervals. On a held-out test set, high fidelity was achieved by single-target regressors (coefficients of determination: Ra 0.9824, Rz 0.9847, RONt 0.9918), with two uncertainty targets also well modelled (standard uncertainty of Ra 0.9899, standard uncertainty of Rz 0.9955); the standard uncertainty of RONt remained more difficult to learn (0.4934). The classifier reached 92.85% accuracy, and probability calibration was essentially unchanged after temperature scaling (expected calibration error 0.00504 → 0.00503 on the test split). Negative transfer was observed for naive multi-output trunks, with single-target models performing better. These results provide calibrated predictions suitable for informing instrument selection and acceptance decisions in metrological workflows.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Division of Metrology and Measurement Systems, Institute of Mechanical Technology, Faculty of Mechanical Engineering, Poznan University of Technology, 60-965 Poznan, Poland.
- Laboratory of Coordinate Metrology, Faculty of Mechanical Engineering, Cracow University of Technology, 31-864 Cracow, Poland.
- Institute of Metrology and Biomedical Engineering, Faculty of Mechatronics, Warsaw University of Technology, 02-525 Warsaw, Poland.
- Department of Metrology and Modern Manufacturing, Faculty of Mechatronics and Mechanical Engineering, Kielce University of Technology, 25-314 Kielce, Poland.
- Central Office of Measures, 00-139 Warsaw, Poland.
- National Information Processing Institute, 00-608 Warsaw, Poland.
ORCID-Profile: Dawid Kucharski, Adam Gąska, Krzysztof Stępień, Bartosz Gapiński, Michal Wieczorowski, Michal Nawotka, Piotr Sobecki, Adam Wójtowicz
Förderung
- Ministry of Education and Science of Poland
Zitieren
APA
Kucharski, D., Gąska, A., Kowaluk, T., Stępień, K., Rępalska, M., Gapiński, B., Wieczorowski, M., Nawotka, M., Sobecki, P., Sosinowski, P., Tomasik, J., & Wójtowicz, A. (2025). Multi-Task Deep Learning for Surface Metrology. Sensors, 25(24), 7471. https://doi.org/10.3390/s25247471
Vancouver
Kucharski D, Gąska A, Kowaluk T, Stępień K, Rępalska M, Gapiński B, et al. Multi-Task Deep Learning for Surface Metrology. Sensors. 2025;25(24):7471. doi:10.3390/s25247471
BibTeX
@article{Kucharski2025,
author = {Kucharski, Dawid and Gąska, Adam and Kowaluk, Tomasz and Stępień, Krzysztof and Rępalska, Marta and Gapiński, Bartosz and Wieczorowski, Michal and Nawotka, Michal and Sobecki, Piotr and Sosinowski, Piotr and Tomasik, Jan and Wójtowicz, Adam},
title = {Multi-Task Deep Learning for Surface Metrology},
journal = {Sensors},
year = {2025},
volume = {25},
number = {24},
pages = {7471},
doi = {10.3390/s25247471}
}
EUVIMED