Abstract
Microstructural analysis and classification are important to understand the microstructure-property relationship of ultrahigh strength steel. Handling this manually takes time, and it requires years of training to reduce the number of errors. Automatic microstructure recognition is driven by machine learning and depends heavily on feature extraction from images at different phases. This paper compared two feature extraction methods, the gray-level cooccurrence matrix (GLCM) and convolutional neural networks (CNNs), to extract the features of different microstructures. The methods were applied to a database of etched and scanned electron microscopy (SEM)-imaged microstructures of ultrahigh strength steel. Driven by metallographic knowledge, each characteristic parameter of the GLCM was analyzed. Then, the t-distributed stochastic neighbor embedding (t-SNE) method was used to compare the two high-dimensional features, and the results showed that feature clustering of the CNN method was better than the GLCM. Subsequently, the features of microstructures were classified by machine learning (ML), and the results showed that the features from CNNs had a better recognition accuracy (98.50%) than those from support vector machine (SVM).
Similar content being viewed by others
References
B. Zhu, W.K. Liang, and Z.X. Gui, J. Metall. Mater. Trans. A 48, 1375–1382. https://doi.org/10.1007/s11661-016-3884-x (2017).
H. Karbasian and A.E. Tekkaya, J. Mater. Process. Technol. 210, 2103–2118. https://doi.org/10.1016/j.jmatprotec.2010.07.019 (2010).
M. Merklein, M. Wieland, and M. Lechner, J. Mater. Process. Technol. 228, 11–24. https://doi.org/10.1016/j.jmatprotec.2015.09.023 (2016).
Y. Mu, B. Wang, and J. Zhou, J. Metall. Mater. Trans. A 48, 5467–5479. https://doi.org/10.1007/s11661-017-4270-z (2017).
B. Zhu, J. Zhu, and Y. Wan, J. Mater. Process. Tech. 262, 392–402. https://doi.org/10.1016/j.jmatprotec.2018.07.011 (2018).
A. Chowdhury, E. Kautz, and B. Yener, Comp. Mater. Sci. 123, 176–187. https://doi.org/10.1016/j.commatsci.2016.05.034 (2016).
S.M. Azimi, D. Britz, and M. Engstler, Sci. Rep-UK 8, 2128. https://doi.org/10.1038/s41598-018-20037-5 (2018).
T. Dutta, D. Das, and S. Banerjee, Measurement 137, 595–603. https://doi.org/10.1016/j.measurement.2018.12.106 (2019).
K. Tsutsui, H. Terasaki, and T. Maemura, Comp. Mater. Sci. 159, 403–411. https://doi.org/10.1016/j.commatsci.2018.12.003 (2019).
J. Gola, D. Britz, and T. Staudt, Comp. Mater. Sci. 148, 324–335. https://doi.org/10.1016/j.commatsci.2018.03.004 (2018).
S. Dutta, K. Barat, and A. Das, Measurement 47, 130–144. https://doi.org/10.1016/j.measurement.2013.08.030 (2014).
B.L. Decost and E.A. Holm, Data Brief 9, 727–731. https://doi.org/10.1016/j.dib.2016.10.011 (2016).
M K F Jr, W G Wee: IEEE 1991 International Conference on Systems Engineering. Dayton, OH, USA, 1991. https://doi.org/10.1109/ICSYSE.1991.161149
A Gebejes, R Huertas, A Tremeau: Color and Imaging Conference, 24th Color and Imaging Conference, 2016, pp. 271-277. https://doi.org/10.2352/ISSN.2169-2629.2017.32.271
T Kobayashi: International Conference on Computer Analysis of Images and Patterns 2015, pp. 594-605. https://doi.org/10.1109/ICSYSE.1991.161149
M. Adankon and M. Cheriet, J. Comput. Sci. 1, 1–28. https://doi.org/10.1007/978-3-642-27733-7_299-3 (2002).
V. Svetnik, J. Chem. Inf. Comput. Sci 43, 133–146. https://doi.org/10.1021/ci034160g (2003).
U. Maulik and S. Bandyopadhyay, J. Pattern Recogn. 33, 1455–1465. https://doi.org/10.1016/S0031-3203(99)00137-5 (2000).
J.P. Yun, S. Choi, and J. Kim, J. NDT E International 42, 389–397. https://doi.org/10.1016/j.ndteint.2009.01.007 (2009).
P. Brynolfsson, D. Nilsson, and T. Torheim, J. Sci Rep 7, 4041. https://doi.org/10.1038/s41598-017-04151-4 (2017).
J. Webel, J. Gola, and D. Britz, J. Mater. Charact. 144, 584–596. https://doi.org/10.1016/j.matchar.2018.08.009 (2018).
J. Gola, D. Britz, and T. Staudt, J. Comput. Mater. Sci. 148, 324–335. https://doi.org/10.1016/j.commatsci.2018.03.004 (2018).
B.L. DeCost and E.A. Holm, J. Comput. Mater. Sci. 110, 126–133. https://doi.org/10.1016/j.commatsci.2015.08.011 (2015).
J Masci, U Meier, D Ciresan: The 2012 International Joint Conference on Neural Networks, 2012, Brisbane, QLD, Australia. https://doi.org/10.1109/IJCNN.2012.6252468
S.M. Azimi, D. Britz, and M. Engstler, J. Sci Rep 8, 2128. https://doi.org/10.1038/s41598-018-20037-5 (2017).
B.L. DeCost, T. Francis, and E.A. Holm, J. Acta Mater. 133, 30–40. https://doi.org/10.1016/j.actamat.2017.05.014 (2017).
N. Lubbers, T. Lookman, and K. Barros, J. Phys. Rev. 96, 052. https://doi.org/10.1103/PhysRevE.96.052111 (2017).
A. Abdollahpoor, X. Chen, and M.P. Pereira, J. Mater. Process. Technol. 228, 125–136. https://doi.org/10.1016/j.jmatprotec.2014.11.033 (2016).
X. Liu: Microstructural characterization of pearlitic and complex phase steels using image analysis methods, University of Birmingham, 2014. http://etheses.bham.ac.uk/id/eprint/4842
R.M.S.K. Haralick, J. IEEE Trans. Syst. Man Cyber. 3, 610–621. https://doi.org/10.1109/TSMC.1973.4309314 (1973).
C. Kavitha, M.B. Rao, and B.P. Rao, Inter. J. Comp. Sci. Inform. Technol. https://doi.org/10.1093/comjnl/bxq066 (2011).
H. Goh, N. Thome, and M. Cord, J. IEEE Trans. Neur. Networks Learn Syst. 25, 2212–2225. https://doi.org/10.1109/TNNLS.2014.2307532 (2017).
K. Gopalakrishnan, S.K. Khaitan and A. Choudhary, Constr. Build. Mater. 157, 322–330. https://doi.org/10.1016/j.conbuildmat.2017.09.110 (2017).
K Simonyan, A Zisserman: J. Comput. Sci., 2014. https://arxiv.org/abs/1409.1556
V D M Laurens, G Hinton: J. Mach. Learn. Res. h, 2008, vlol.9, pp. 2579-2605. http://www.cs.toronto.edu/~hinton/absps/tsne.pdf
Acknowledgements
This research work was financially supported by the National Natural Science Foundation of China (Grant No. U1760205) and the National Science and Technology Major Project of China (Grant No. 2018ZX04023001).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
We declare that we have no financial and personal relationships with people or organizations that would inappropriately influence our work, and there is no professional or other personal interest of any nature in any product, service, and/or company that could influence either the position presented in or the review of this manuscript.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Zhu, B., Chen, Z., Hu, F. et al. Feature Extraction and Microstructural Classification of Hot Stamping Ultra-High Strength Steel by Machine Learning. JOM 74, 3466–3477 (2022). https://doi.org/10.1007/s11837-022-05265-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11837-022-05265-5