journal article·2024-03Open access

A new paradigm in electron microscopy: Automated microstructure analysis utilizing a dynamic segmentation convolutional neutral network

Stephen Taller, Luke Scime, Ty Austin

Abstract

Over the past half century, the transmission electron microscope enabled insight into the fundamental arrangements and structures of materials. State-of-the-art electron microscopes can acquire large image datasets across multiple imaging modalities. However, the manual annotation process for feature or defect quantification may not be feasible with the modern microscope. Convolutional neural networks emerged to characterize individual microstructural features from an image in a cost-effective, consistent manner. However, many of these neural network approaches rely on thousands to hundreds of thousands of manual annotations of each feature type across hundreds of images to train the network for adequate performance. This work focused on the development and application of a pixel-wise defect detection machine-learning dynamic segmentation convolutional neural network with associated automated acquisition and postprocessing to identify microstructural features rapidly and quantitatively from a small initial dataset incorporating multiple imaging modes. The approach was demonstrated for characterization of superalloy 718 from both single image acquisition on multiple detectors to in-situ evolution captured with a single detector on a standard desktop computer to demonstrate the low barrier to entry required for widespread adoption. Pixel-by-pixel class identification was excellent with strong identification of chemically distinct phases, structurally distinct phases, and defect structures, thus demonstrating the new paradigm of machine learning-assisted characterization.
Source: openalex_inverted_index · Quality: baseline

Research context

OpenAlex topics: Machine Learning in Materials Science · Electron and X-Ray Spectroscopy Techniques · Nuclear Physics and Applications

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

References (30)

Show source-provided references
1. Automated defect analysis in electron microscopic images10.1038/s41524-018-0093-8 ↗
2. Multi defect detection and analysis of electron microscopy images with deep learning10.1016/j.commatsci.2021.110576 ↗
3. Automated detection of Helium bubbles in irradiated X-75010.1016/j.ultramic.2020.113068 ↗
4. Misra · 2020 · 289
5. A deep learning based automatic defect analysis framework for In-situ TEM ion irradiations10.1016/j.commatsci.2021.110560 ↗
6. In-situ TEM investigation of void swelling in Nickel under irradiation with analysis aided by computer vision10.1016/j.actamat.2023.119013 ↗
7. Oradei-Basile · A Current T-T-T Diagram for Wrought Alloy 718 · 2012 · 325
8. Toward a better understanding of phase transformations in additive manufacturing of Alloy 71810.1016/j.mtla.2020.100862 ↗
9. Radavich · 2004 · 229
10. Taller · 2021
11. Using post-processing heat treatments to elucidate precipitate strengthening of additively manufactured superalloy 718
12. Utilizing a dynamic segmentation convolutional neural network for microstructure analysis of additively manufactured superalloy 71810.1017/S143192762101076X ↗
13. Specification · 2010 · vol. 7 · 7
14. Application of STEM characterization for investigating radiation effects in BCC Fe-based alloys10.1557/jmr.2015.32 ↗
15. STEM characterization of dislocation loops in irradiated FCC alloys10.1016/j.jnucmat.2020.152658 ↗
16. TEM characterization of dislocation loops in irradiated bcc Fe-based steels10.1016/j.jnucmat.2012.12.002 ↗
17. Phase-contrast imaging in aberration-corrected scanning transmission electron microscopy10.1016/j.micron.2013.03.006 ↗
18. Peña · hyperspy/hyperspy: Release v1.7.1 · 2022
19. The quantitative analysis of thin specimens10.1111/j.1365-2818.1975.tb03895.x ↗
20. An FFT-based technique for translation, rotation, and scale-invariant image registration10.1109/83.506761 ↗
21. C. Gohlke and M. Týč, "imreg_dft," ed: GitHub Repository ..
22. Layer-wise anomaly detection and classification for powder bed additive manufacturing processes: a machine-agnostic algorithm for real-time pixel-wise semantic segmentation
23. "U-Net: convolutional networks for biomedical image segmentation," in international conference on medical image computing and computer-assisted intervention
24. Simple online and realtime tracking
25. Understanding bubble and void nucleation in dual ion irradiated T91 steel using single parameter experiments10.1016/j.actamat.2020.07.060 ↗
26. Dislocation loop evolution during in-situ ion irradiation of model FeCrAl alloys10.1016/j.actamat.2017.07.011 ↗
27. Effect of dose on irradiation-induced loop density and Burgers vector in ion-irradiated ferritic/martensitic steel HT910.1080/14786435.2018.1490825 ↗
28. Characterization of faulted dislocation loops and cavities in ion irradiated alloy 800H10.1016/j.jnucmat.2017.11.012 ↗
29. Al-Najjar · 2022
30. The INTERSECT open federated architecture for the laboratory of the future