Machine-Learning-Based Evaluation of Ferrite Grain Size from Microstructural Images
Journal of the Japan Institute of Metals and Materials · 01.11.2026 · 90 (11) · 209-215
DOI: 10.2320/jinstmet.j202620 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of the Japan Institute of Metals and Materials, erschienen am 01.11.2026. Verfasst von Taketo Yamada und 2 weiteren Autor:innen. Die Arbeit nennt 20 Literaturangaben.
Abstract (Originalsprache)
A deep-learning approach is proposed for automatic measurement of ferrite grain size in ferrite–pearlite steel by the intercept method specified in JIS G 0551. Unlike conventional image-analysis methods that estimate grain size from segmented areas, the proposed method directly classifies each pixel on a one-pixel-wide horizontal test line into six classes, reproducing the manual intercept procedure. SAM2-UNet, which combines the pre-trained encoder of SAM2 with a U-Net decoder, is used as the segmentation model. By explicitly training the model to recognize etch pits, polishing scratches, and inclusions, confusion with grain boundaries is suppressed. A composite loss function of weighted cross-entropy and Dice loss, together with an edge-weight mask, is introduced. For 27 specimens of case-hardened steel, the proposed method achieved a correlation coefficient of r = 0.946 with manual measurement and a mean absolute percentage error of 9.39%.
Zitieren
APA
Yamada, T., Ono, R., & Murase, H. (2026). Machine-Learning-Based Evaluation of Ferrite Grain Size from Microstructural Images. Journal of the Japan Institute of Metals and Materials, 90(11), 209-215. https://doi.org/10.2320/jinstmet.j202620
Vancouver
Yamada T, Ono R, Murase H. Machine-Learning-Based Evaluation of Ferrite Grain Size from Microstructural Images. Journal of the Japan Institute of Metals and Materials. 2026;90(11):209-215. doi:10.2320/jinstmet.j202620
BibTeX
@article{Yamada2026,
author = {Yamada, Taketo and Ono, Ryonosuke and Murase, Hironori},
title = {Machine-Learning-Based Evaluation of Ferrite Grain Size from Microstructural Images},
journal = {Journal of the Japan Institute of Metals and Materials},
year = {2026},
volume = {90},
number = {11},
pages = {209--215},
doi = {10.2320/jinstmet.j202620}
}
EUVIMED