Pith. sign in

REVIEW 4 major objections 4 minor 73 references

Towards Space Group Determination from EBSD Patterns: The Role of Deep Learning and High-throughput Dynamical Simulations

T0 review · 4 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Neural networks can classify cubic space groups from single EBSD patterns once the training target is the compositionally disordered structure, reaching 98 percent simulated and over 90 percent experimental accuracy.

desk verdict Relabeling idea and the 5,148-phase simulated dataset are the real contributions; the 90% experimental accuracy claim is not supported because test phases are in training and the labels themselves are inherited from a dataset the paper shows to be unreliable. read the letter →

arxiv 2504.21331 v2 pith:BRCXP37U submitted 2025-04-30 cond-mat.mtrl-sci cs.CV

Alfred Yan , Muhammad Nur Talha Kilic , Gert Nolze , Ankit Agrawal , Alok Choudhary , Roberto dos Reis , Vinayak Dravid This is my paper · ORCID
classification cond-mat.mtrl-scics.CV
keywords deeplearningEBSDsimulationsymmetryspacegroupdomainadaptationelectronmicroscopy
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that a neural network can classify the cubic space group of a material directly from a single electron backscatter diffraction (EBSD) pattern, provided the prediction target is the space group of the equivalent compositionally disordered structure rather than the true, chemically ordered space group. On simulated patterns from 5,148 cubic phases, the relabeling scheme raises cross-validation accuracy from 91 percent to 98 percent. On experimental patterns from the same phases, unsupervised domain adaptation yields 0.71 average accuracy overall and 0.89 after excluding two low-pattern-quality phases, with ensemble voting reaching 0.93. If this holds, EBSD plus deep learning could become a fast, scalable symmetry screen in high-throughput materials discovery.

What carries the argument

The load-bearing mechanism is a relabeling scheme: for each crystal structure, the label presented to the network is not the true space group but the space group of the compositionally disordered equivalent, obtained by replacing every atom in the unit cell with the same element. This converts the task from detecting subtle chemical-ordering superstructure reflections, which are often too weak or absent in EBSD patterns, to identifying the underlying elemental lattice and its three-dimensional symmetry, which dynamical simulations reproduce faithfully. The scheme is paired with high-throughput Bloch-wave dynamical simulations of 5,148 cubic phases and with Maximum Classifier Discrepancy, an adversarial unsupervised domain adaptation method that aligns simulated and experimental pattern distributions.

What would settle it

Run an independent diffraction experiment with sensitivity to chemical ordering, such as synchrotron powder X-ray diffraction or transmission electron diffraction, on the same retained-phase samples used here. If any retained phase is found to have a space group different from the label assigned in this paper, or if the relabeled Im-3m assignment for Al4CoNi2 is wrong, the reported 0.71 to 0.93 experimental accuracies do not measure space-group classification as claimed.

Watch

Extended reading notes

Core claim

The central discovery is that the main obstacle to space-group classification from EBSD patterns is not the network's ability to read symmetry from Kikuchi bands, but the choice of label: compositional ordering produces pseudosymmetry, because patterns from structures that differ only by chemical ordering can be nearly identical when atomic scattering factors are similar. The paper shows that retargeting the classifier to the space group obtained after setting all atoms to the same element removes most of that ambiguity. With this relabeling, a residual convolutional network trained and tested on disjoint sets of simulated phases reaches 98 percent cross-validation accuracy, and a Maximum Classifier Discrepancy domain-adaptation model trained jointly on simulated and unlabeled experimental patterns classifies experimental EBSD patterns of the same phases with better than 90 percent accuracy after low-quality patterns are removed. The paper also uses manual lattice analysis to show that previously assigned Ia-3d labels in the experimental dataset were unreliable, and it relabels the Al4CoNi2 phase as Im-3m accordingly.

Load-bearing premise

The experimental accuracy numbers rest on the assumption that the phase labels inherited from the prior experimental dataset are correct for the five retained space-group classes; the paper's own audit shows those labels were wrong for at least one phase, Al4Ni3, and no independent check is reported for the rest.

Editorial extensions

If this is right

  • A model trained on the relabeled scheme can classify an EBSD pattern's cubic space-group type without being told which phase is present, so no prior phase library is needed at prediction time.
  • Phases that share a Bravais lattice but sit in different space groups, such as Pm-3m versus Im-3m B2-type structures, become separable because the target is the underlying elemental lattice.
  • Accuracy on a phase is tied to the quality of its experimental patterns rather than to the phase's intrinsic difficulty: NiAl and Al are classified perfectly in simulation but poorly in experiment.
  • With compositional information supplied separately, the predicted disordered space group narrows the candidate crystal structures and supports high-throughput structure determination.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Inference: If the relabeling strategy transfers to other crystal systems, EBSD-based structure determination could split into two steps: classify the disordered parent lattice from the pattern, then determine chemical ordering from composition and separate measurements.
  • Inference: The 98 percent figure comes from simulated patterns with high band contrast; real-world pattern quality, as the NiAl and Al results show, is likely to cap the accuracy unless acquisition is optimized or low-quality patterns are filtered.
  • Inference: The method's real test is generalization to phases never seen in training; the paper's own experimental evaluation uses the same phases as training, so an immediate extension is to apply the trained ensemble to new cubic phases and measure accuracy.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper proposes a deep-learning framework for classifying cubic space groups from electron backscatter diffraction (EBSD) patterns. A large simulated dataset of 5,148 cubic phases is created with EMsoft dynamical simulations. The authors first train a ResNet18 on simulated patterns to predict the space group of the compositionally disordered equivalent structure, reporting 98% phase-disjoint cross-validation accuracy versus 91% when predicting the true space group. They then use Maximum Classifier Discrepancy (MCD) domain adaptation to train on simulated plus unlabeled experimental patterns from Kaufmann et al. For experimental patterns, they report 0.71 average accuracy over all phases, rising to 0.89–0.93 after removing NiAl and Al and applying ensemble voting, with the target labels also relabeled to the compositionally disordered space groups. The central claim is that this relabeling scheme enables accuracies above 90% on both simulated and experimental data, indicating neural networks can extract symmetry information from EBSD patterns.

Significance. The simulated cross-validation part is a solid contribution: it uses phase-disjoint folds, class-balanced sampling, and a physically motivated relabeling target, showing that a network can distinguish five cubic space-group types from simulated patterns with high accuracy. The paper also usefully demonstrates that the largest confusion source in true-space-group classification is compositional ordering, and its quantitative comparison (91% vs 98%) supports that interpretation. However, the experimental validation does not support the abstract's broad claim: the high accuracy is obtained only after excluding two phases and only on test patterns from the same phases used in training, and the ground-truth labels are inherited without independent verification. If the simulated results were the main claim, this would be a solid methods paper; as written, the experimental 'towards space group determination' claim needs substantial qualification.

major comments (4)
  1. [Abstract and Results (Figure 4a)] The abstract claims 'accuracy scores higher than 90% on simulated and experimental data,' but the experimental accuracy is 0.71±0.01 for all phases and only reaches 0.89±0.03 average (0.93 ensemble) after removing NiAl and Al (Figure 4a). The reported 90%+ experimental accuracy is thus conditional on excluding two of the seventeen phases, and this conditional nature should be stated wherever the 90% figure appears.
  2. [Results, CALM analysis and Table 1] The experimental evaluation inherits the space-group labels from Kaufmann et al. for the five retained space-group types (Pm-3m, Pm-3n, Fm-3m, Fd-3m, Im-3m). The paper's own CALM audit of the Ia-3d class shows that such inherited labels are unreliable: Al4Ni3 patterns fitted a hexagonal/trigonal lattice and Al4CoNi2 patterns fitted a primitive lattice, leading the authors to relabel or exclude them. No comparable independent check is reported for the retained classes, yet the reported experimental accuracies are computed against those labels. If any retained phase carries an incorrect label (e.g., NiAl may not be fully B2-ordered), the stated accuracies do not measure space-group classification. Because the experimental support of the central claim rests entirely on these labels, this is a load-bearing weakness.
  3. [Summary and Outlook (final paragraph)] The paper explicitly acknowledges that 'the models here are trained and tested on the same phases–an upcoming version will present results showing the model performance on novel phases outside the training dataset.' This means the experimental results demonstrate within-distribution classification, not the ability to determine the space group of an unknown phase, which is the stated motivation (megalibraries, unknown samples). The abstract and introduction should be tempered to reflect that the experimental claims are limited to well-characterized phases whose patterns are seen during training.
  4. [Results, relabeling scheme] The model predicts the space group of the compositionally disordered equivalent structure, not the true space group of the material. The paper argues that composition, available separately, can resolve the ambiguity, but no demonstration or algorithm is provided for combining the network's disordered-space-group prediction with composition to recover the true space group. The central phrase 'predictions of crystal symmetry from an EBSD pattern' overstates what is demonstrated; the network output is a symmetry label of a hypothetical disordered structure, and the mapping to the actual crystal symmetry is left as future work.
minor comments (4)
  1. [Methods, MCD training] The Results section states that MCD models were trained for 30 runs, while SI Figure 7 reports averages 'over 20 runs.' The discrepancy should be resolved or clarified.
  2. [General formatting] Several space-group symbols are rendered with spaces, e.g., 'Pm m' instead of 'Pm-3m' or 'Pm3m.' Please use standard crystallographic notation throughout, including in Table 1 and the supplementary figures.
  3. [SI Figure references] The text refers to 'Si Figure 3', 'Si Figure 4', 'Si Figure 5', etc.; these should be 'SI Figure' with consistent numbering, and the supplementary figure labels themselves should be checked for matching numbers.
  4. [Methods, simulation parameters] The pattern center offset and pixel size are described as guessed or taken from the literature, and the paper states that a perfect match was not required because domain adaptation would adjust. It would strengthen the paper to state explicitly how sensitive the simulated cross-validation results are to these parameter choices.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity found; the simulated cross-validation, relabeling scheme, and MCD domain adaptation are self-contained, and the same-phase experimental evaluation is disclosed as a limitation rather than disguised as a derivation.

full rationale

The paper's derivation chain is self-contained. Simulated training labels are obtained by applying compositional disordering to Materials Project crystal structures, a well-defined structural transformation that does not depend on the network's outputs. The 5-fold cross-validation on simulated patterns uses disjoint sets of phases, so the reported 98% accuracy is a genuine generalization test on unseen phases, not a fitted-input prediction. The relabeling scheme is a deliberate change of the target variable, and the paper explicitly compares it with true-space-group prediction; the improvement is an empirical finding, not a definitional tautology. The MCD experimental evaluation is trained and tested on the same phases, but the paper states this directly in the final section: 'models here are trained and tested on the same phases–an upcoming version will present results showing the model performance on novel phases outside the training dataset.' That is a disclosed scope limitation, not a circular derivation. The inherited experimental space-group labels from Kaufmann et al. are external ground-truth inputs; their reliability is a correctness concern, and the paper even audits one class with CALM and revises labels accordingly. This shows responsiveness to label quality rather than a self-referential argument. No load-bearing step reduces by construction to its own inputs, and no uniqueness claim is imported from the authors' prior work.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The pipeline rests on the simulation being a faithful proxy for real EBSD patterns, on the correctness of the structural database, and on the relabeling scheme being the right target. The guessed pattern center and pixel size, the volume cutoffs, and the B-factor range are hand-set simulation parameters; none are fitted to the space group labels, but they shape the training distribution. No invented entities are introduced.

free parameters (6)
  • Simulated pattern center offset = 171 microns above detector center
    Chosen from the literature (ref 68), not measured for the experimental setup; used for all gnomonic projection simulations. The paper states it was an arbitrary value.
  • Simulated pixel size = 23 microns
    Guessed by the authors, who say it 'by inspection seemed to provide good agreement' with experimental data. A hand-picked simulation parameter.
  • Unit cell volume cutoffs for simulation selection = 250, 500, and 2000 cubic angstroms depending on space group
    Inclusion thresholds that change the training distribution per space group, chosen for computational efficiency rather than scientific completeness.
  • B-factor augmentation range = 0.005 to 0.011 nm^2
    Randomly selected per atom in re-simulations to mimic experimental pattern degradation, especially for Ta. This is a hand-set data augmentation range.
  • Image quality threshold for experimental training sampling = 100 highest IQ images per phase
    The 100 highest quality patterns per phase were sampled for MCD training, which biases the training distribution and likely improves accuracy relative to a random sample.
  • MCD training epochs and discrepancy iterations = 20 epochs; n=4
    Hyperparameters taken from the original MCD study, not tuned on this dataset, but they directly affect the reported experimental accuracy.
assumptions (6)
  • domain assumption Dynamical diffraction theory (Bloch wave approach) as implemented in EMsoft accurately reproduces EBSD pattern intensities for the simulated structures.
    The entire simulated dataset is generated by EMsoft; if the simulations miss physics important for symmetry, the classifier learns the wrong cues. Invoked in the Methods simulation section.
  • domain assumption Materials Project CIF files provide correct and representative crystal structures for the selected cubic phases.
    Simulated patterns are computed from these structures; wrong structures would give wrong labels. Invoked in Methods: High-throughput simulation.
  • ad hoc to paper The space group of the compositionally disordered equivalent structure is a well-defined and learnable target that, combined with composition, determines the true space group.
    The relabeling scheme converts the problem from true space group to disordered-space group. This is a new modeling choice central to the claimed high accuracy; it is motivated by pseudosymmetry but not proven to always yield the true space group.
  • domain assumption The experimental ground truth labels from Kaufmann et al. are correct for the five retained space group types.
    Used as test labels in Figure 4; the paper audited only Ia-3d patterns with CALM and found one phase mislabeled, so the retained labels are an unverified inheritance.
  • standard math MCD assumes that source and target domains share the same label set, and this holds for the five retained classes.
    Standard condition for unsupervised domain adaptation; the paper excludes Ia-3d partly to maintain this. Invoked in Methods: MCD model.
  • domain assumption ImageNet-pretrained ResNet features transfer to EBSD pattern classification.
    Models are initialized with ImageNet weights; this is a common but unproven assumption for non-natural images and could affect accuracy.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Space Group Determination from EBSD Patterns: The Role of Deep Learning and High-throughput Dynamical Simulations." pith.science (2026). https://pith.science/paper/BRCXP37U

@misc{pith2026250421331,
  author       = {Pith},
  title        = {Pith review of: Towards Space Group Determination from EBSD Patterns: The Role of Deep Learning and High-throughput Dynamical Simulations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BRCXP37U}},
  note         = {Machine review of arXiv:2504.21331}
}
read the original abstract

The design of novel materials hinges on the understanding of structure-property relationships. However, in recent times, our capability to synthesize a large number of materials has outpaced our speed at characterizing them. While the overall chemical constituents can be readily known during synthesis, the structural evolution and characterization of newly synthesized samples remains a bottleneck for the ultimate goal of high throughput nanomaterials discovery. Thus, scalable methods for crystal symmetry determination that can analyze a large volume of material samples within a short time-frame are especially needed. Kikuchi diffraction in the SEM is a promising technique for this due to its sensitivity to dynamical scattering, which may provide information beyond just the seven crystal systems and fourteen Bravais lattices. After diffraction patterns are collected from material samples, deep learning methods may be able to classify the space group symmetries using the patterns as input, which paired with the elemental composition, would help enable the determination of the crystal structure. To investigate the feasibility of this solution, neural networks were trained to predict the space group type of background corrected EBSD patterns. Our networks were first trained and tested on an artificial dataset of EBSD patterns of 5,148 different cubic phases, created through physics-based dynamical simulations. Next, Maximum Classifier Discrepancy, an unsupervised deep learning-based domain adaptation method, was utilized to train neural networks to make predictions for experimental EBSD patterns. We introduce a relabeling scheme, which enables our models to achieve accuracy scores higher than 90% on simulated and experimental data, suggesting that neural networks are capable of making predictions of crystal symmetry from an EBSD pattern.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

73 extracted references · 67 canonical work pages

  1. [1]

    Kluender, E. J. et al. Catalyst discovery through megalibraries of nanomaterials. Proceedings of the National Academy of Sciences 116 , 40–45 (2019)

    work page 2019

  2. [2]

    Huo, F. et al. Polymer Pen Lithography. Science 321 , 1658–1660 (2008)

    work page 2008

  3. [3]

    B., Swisher, J

    Wahl, C. B., Swisher, J. H., Smith, P. T., Dravid, V. P. & Mirkin, C. A. Traversing the Periodic Table through Phase-Separating Nanoreactors. Advanced Materials n/a , 2500088

    work page

  4. [4]

    A., Roth, E

    Hujsak, K. A., Roth, E. W., Kellogg, W., Li, Y. & Dravid, V. P. High speed/low dose analytical electron microscopy with dynamic sampling. Micron 108 , 31–40 (2018)

    work page 2018

  5. [5]

    B., Mirkin, C

    Wahl, C. B., Mirkin, C. A. & Dravid, V. P. Towards Autonomous Electron Microscopy for High-throughput Materials Discovery. Microscopy and Microanalysis 29 , 1913–1914 (2023)

    work page 2023

  6. [6]

    Smith, P. T. et al. Molecular Thin Films Enable the Synthesis and Screening of Nanoparticle Megalibraries Containing Millions of Catalysts. J. Am. Chem. Soc. 145 , 14031–14043 (2023)

    work page 2023

  7. [7]

    Day, A. L. et al. Machine Learning-Enabled Image Classification for Automated Electron Microscopy. Microscopy and Microanalysis 30 , 456–465 (2024)

    work page 2024

  8. [8]

    Ra, M. et al. Classification of crystal structures using electron diffraction patterns with a deep convolutional neural network. RSC Advances 11 , 38307–38315 (2021)

    work page 2021

Show all 73 references
  1. [9]

    & Aitkaliyeva, A

    Mika, M., Tomczak, N., Finney, C., Carter, J. & Aitkaliyeva, A. Automating selective area electron diffraction phase identification using machine learning. Journal of Materiomics 10 , 896–905 (2024)

    2024

  2. [10]

    & Ghiringhelli, L

    Ziletti, A., Kumar, D., Scheffler, M. & Ghiringhelli, L. M. Insightful classification of crystal structures using deep learning. Nat Commun 9 , 2775 (2018)

    2018

  3. [11]

    Tiong, L. C. O., Kim, J., Han, S. S. & Kim, D. Identification of crystal symmetry from noisy diffraction patterns by a shape analysis and deep learning. npj Comput Mater 6 , 1–11 (2020)

    2020

  4. [12]

    Chen, J. et al. Automated crystal system identification from electron diffraction patterns using multiview opinion fusion machine learning. Proceedings of the National Academy of Sciences 120 , e2309240120 (2023)

    2023

  5. [13]

    A., Gong, M

    Aguiar, J. A., Gong, M. L., Unocic, R. R., Tasdizen, T. & Miller, B. D. Decoding crystallography from high-resolution electron imaging and diffraction datasets with deep learning. Science Advances 5 , eaaw1949 (2019)

    2019

  6. [14]

    Gleason, S. P. et al. Random forest prediction of crystal structure from electron diffraction patterns incorporating multiple scattering. Phys. Rev. Mater. 8 , 093802 (2024)

    2024

  7. [15]

    E., Lerman, S., Du, Z., Xu, C

    Salgado, J. E., Lerman, S., Du, Z., Xu, C. & Abdolrahim, N. Automated classification of big X-ray diffraction data using deep learning models. npj Comput Mater 9 , 1–12 (2023)

    2023

  8. [16]

    M., Choo, K., Chang, J

    Vecsei, P. M., Choo, K., Chang, J. & Neupert, T. Neural network based classification of crystal symmetries from x-ray diffraction patterns. Phys. Rev. B 99 , 245120 (2019)

    2019

  9. [17]

    Niitsu, N. et al. Powder x-ray diffraction analysis with machine learning for organic-semiconductor crystal-structure determination. Applied Physics Letters 125 , 013301 (2024)

    2024

  10. [18]

    Suzuki, Y. et al. Symmetry prediction and knowledge discovery from X-ray diffraction patterns using an interpretable machine learning approach. Sci Rep 10 , 21790 (2020)

    2020

  11. [19]

    Chen, L. et al. Crystal Structure Assignment for Unknown Compounds from X-ray Diffraction Patterns with Deep Learning. J. Am. Chem. Soc. 146 , 8098–8109 (2024)

    2024

  12. [20]

    Riesel, E. A. et al. Crystal Structure Determination from Powder Diffraction Patterns with Generative Machine Learning. J. Am. Chem. Soc. 146 , 30340–30348 (2024)

    2024

  13. [21]

    Isabell, T. C. & Dravid, V. P. Resolution and sensitivity of electron backscattered diffraction in a cold field emission gun SEM. Ultramicroscopy 67 , 59–68 (1997)

    1997

  14. [22]

    C., Singh, S

    Lenthe, W. C., Singh, S. & Graef, M. D. A spherical harmonic transform approach to the indexing of electron back-scattered diffraction patterns. Ultramicroscopy 207 , 112841 (2019)

    2019

  15. [23]

    Wright, S. I. & Adams, B. L. Automatic analysis of electron backscatter diffraction patterns. Metall Trans A 23 , 759–767 (1992)

    1992

  16. [24]

    Wright, S. I. Fundamentals of Automated EBSD. in Electron Backscatter Diffraction in Materials Science (eds. Schwartz, A. J., Kumar, M. & Adams, B. L.) 51–64 (Springer US, Boston, MA, 2000). doi:10.1007/978-1-4757-3205-4_5

    doi

  17. [25]

    & Boyle, A

    Nolze, G., Winkelmann, A. & Boyle, A. P. Pattern matching approach to pseudosymmetry problems in electron backscatter diffraction. Ultramicroscopy 160 , 146–154 (2016)

    2016

  18. [26]

    & Winkelmann, A

    Nolze, G., Jürgens, M., Olbricht, J. & Winkelmann, A. Improving the precision of orientation measurements from technical materials via EBSD pattern matching. Acta Materialia 159 , 408–415 (2018)

    2018

  19. [27]

    & Tokarski, T

    Nolze, G., Winkelmann, A., Cios, G. & Tokarski, T. Tetragonality mapping of martensite in a high ‑ carbon steel by EBSD. Materials Characterization 175 , 111040 (2021)

    2021

  20. [28]

    Chen, Y. H. et al. A Dictionary Approach to Electron Backscatter Diffraction Indexing*. Microscopy and Microanalysis 21 , 739–752 (2015)

    2015

  21. [29]

    & Nolze, G

    Winkelmann, A. & Nolze, G. Chirality determination of quartz crystals using Electron Backscatter Diffraction. Ultramicroscopy 149 , 58–63 (2015)

    2015

  22. [30]

    Burkhardt, U. et al. Absolute Structure from Scanning Electron Microscopy. Sci Rep 10 , 4065 (2020)

    2020

  23. [31]

    & Rüssel, C

    Zscheckel, T., Wisniewski, W. & Rüssel, C. Polarity Determination in EBSD Patterns Using the Hough Transformation. Microscopy and Microanalysis 27 , 1409–1419 (2021)

    2021

  24. [32]

    & Winkelmann, A

    Nolze, G., Grosse, C. & Winkelmann, A. Kikuchi pattern analysis of noncentrosymmetric crystals. J Appl Cryst 48 , 1405–1419 (2015)

    2015

  25. [33]

    Winkelmann, A., Trager-Cowan, C., Sweeney, F., Day, A. P. & Parbrook, P. Many-beam dynamical simulation of electron backscatter diffraction patterns. Ultramicroscopy 107 , 414–421 (2007)

    2007

  26. [34]

    & Fortunier, R

    Maurice, C., Dzieciol, K. & Fortunier, R. A method for accurate localisation of EBSD pattern centres. Ultramicroscopy 111 , 140–148 (2011)

    2011

  27. [35]

    Callahan, P. G. & Graef, M. D. Dynamical Electron Backscatter Diffraction Patterns. Part I: Pattern Simulations. Microscopy and Microanalysis 19 , 1255–1265 (2013)

    2013

  28. [36]

    & Werner, W

    Winkelmann, A., Nolze, G., Vos, M., Salvat-Pujol, F. & Werner, W. S. M. Physics-based simulation models for EBSD: advances and challenges. IOP Conf. Ser.: Mater. Sci. Eng. 109 , 012018 (2016)

    2016

  29. [37]

    J., Kumar, A

    Shen, Y.-F., Pokharel, R., Nizolek, T. J., Kumar, A. & Lookman, T. Convolutional neural network-based method for real-time orientation indexing of measured electron backscatter diffraction patterns. Acta Materialia 170 , 118–131 (2019)

    2019

  30. [38]

    Jha, D. et al. Extracting Grain Orientations from EBSD Patterns of Polycrystalline Materials Using Convolutional Neural Networks. Microsc Microanal 24 , 497–502 (2018)

    2018

  31. [39]

    Xiong, G., Wang, C., Yan, Y., Zhang, L. & Su, Y. Indexing high-noise electron backscatter diffraction patterns using convolutional neural network and transfer learning. Computational Materials Science 233 , 112718 (2024)

    2024

  32. [40]

    & De Graef, M

    Ding, Z., Pascal, E. & De Graef, M. Indexing of electron back-scatter diffraction patterns using a convolutional neural network. Acta Materialia 199 , 370–382 (2020)

    2020

  33. [41]

    Jangid, D. K. et al. Q-RBSA: high-resolution 3D EBSD map generation using an efficient quaternion transformer network. npj Comput Mater 10 , 1–10 (2024)

    2024

  34. [42]

    Wanni, J., Bronkhorst, C. A. & Thoma, D. J. Machine learning enhanced analysis of EBSD data for texture representation. npj Comput Mater 10 , 1–11 (2024)

    2024

  35. [43]

    & De Graef, M

    Ding, Z., Zhu, C. & De Graef, M. Determining crystallographic orientation via hybrid convolutional neural network. Materials Characterization 178 , 111213 (2021)

    2021

  36. [44]

    & Vecchio, K

    Kaufmann, K., Lane, H., Liu, X. & Vecchio, K. S. Efficient few-shot machine learning for classification of EBSD patterns. Sci Rep 11 , 8172 (2021)

    2021

  37. [45]

    Kaufmann, K. et al. Phase Mapping in EBSD Using Convolutional Neural Networks. Microscopy and Microanalysis 26 , 458–468 (2020)

    2020

  38. [46]

    Breumier, S. et al. Leveraging EBSD data by deep learning for bainite, ferrite and martensite segmentation. Materials Characterization 186 , 111805 (2022)

    2022

  39. [47]

    Martinez Ostormujof, T. et al. Deep Learning for automated phase segmentation in EBSD maps. A case study in Dual Phase steel microstructures. Materials Characterization 184 , 111638 (2022)

    2022

  40. [48]

    & Winkelmann, A

    Nolze, G., Hielscher, R. & Winkelmann, A. Electron backscatter diffraction beyond the mainstream. Crystal Research and Technology 52 , 1600252 (2017)

    2017

  41. [49]

    & Zhu, Y

    Han, M., Zhao, G. & Zhu, Y. Accurate determination of low-symmetry Bravais unit cells by EBSD. Ultramicroscopy 195 , 136–146 (2018)

    2018

  42. [50]

    & Winkelmann, A

    Nolze, G., Tokarski, T., Rychłowski, Ł., Cios, G. & Winkelmann, A. Crystallographic analysis of the lattice metric (CALM) from single electron backscatter diffraction or transmission Kikuchi diffraction patterns. J Appl Cryst 54 , 1012–1022 (2021)

    2021

  43. [51]

    Dingley, D. J. & Wright, S. I. Determination of crystal phase from an electron backscatter diffraction pattern. Journal of Applied Crystallography 42 , 234–241 (2009)

    2009

  44. [52]

    & Han, M

    Li, L. & Han, M. Determining the Bravais lattice using a single electron backscatter diffraction pattern. J Appl Crystallogr 48 , 107–115 (2015)

    2015

  45. [53]

    Dingley, D. J. & Baba-Kishi, K. Use of Electron Back Scatter Diffraction Patterns for Determination of Crystal Symmetry Elements. Scanning Electron Microscopy 1986 , 383–391

    1986

  46. [54]

    & Rychłowski, Ł

    Nolze, G., Tokarski, T. & Rychłowski, Ł. Use of electron backscatter diffraction patterns to determine the crystal lattice. Part 3. Pseudosymmetry. J Appl Crystallogr 56 , 367–380 (2023)

    2023

  47. [55]

    & Westphal, C

    Lühr, T., Winkelmann, A., Nolze, G., Krull, D. & Westphal, C. Direct Atom Imaging by Chemical-Sensitive Holography. Nano Lett. 16 , 3195–3201 (2016)

    2016

  48. [56]

    Baba-Kishi, K. Z. Review Electron backscatter Kikuchi diffraction in the scanning electron microscope for crystallographic analysis. Journal of Materials Science 37 , 1715–1746 (2002)

    2002

  49. [57]

    Kaufmann, K. et al. Crystal symmetry determination in electron diffraction using machine learning. Science 367 , 564–568 (2020)

    2020

  50. [58]

    Kaufmann, K., Zhu, C., Rosengarten, A. S. & Vecchio, K. S. Deep Neural Network Enabled Space Group Identification in EBSD. Microscopy and Microanalysis 26 , 447–457 (2020)

    2020

  51. [59]

    Jain, A. et al. Commentary: The Materials Project: A materials genome approach to accelerating materials innovation. APL Materials 1 , 011002 (2013)

    2013

  52. [60]

    Liu, X. et al. Deep Unsupervised Domain Adaptation: A Review of Recent Advances and Perspectives. Preprint at http://arxiv.org/abs/2208.07422 (2022)

    2022 arXiv

  53. [61]

    & De Graef, M

    Singh, S., Ram, F. & De Graef, M. EMsoft: open source software for electron diffraction/image simulations. Microscopy and Microanalysis 23 , 212–213 (2017)

    2017

  54. [62]

    & Sun, J

    He, K., Zhang, X., Ren, S. & Sun, J. Deep Residual Learning for Image Recognition. Preprint at https://doi.org/10.48550/arXiv.1512.03385 (2015)

    2015 doi arXiv

  55. [63]

    Ivantchev, S., Kroumova, E., Madariaga, G., Pérez-Mato, J. M. & Aroyo, M. I. SUBGROUPGRAPH: a computer program for analysis of group–subgroup relations between space groups. J Appl Cryst 33 , 1190–1191 (2000)

    2000

  56. [64]

    & Harada, T

    Saito, K., Watanabe, K., Ushiku, Y. & Harada, T. Maximum Classifier Discrepancy for Unsupervised Domain Adaptation. Preprint at http://arxiv.org/abs/1712.02560 (2018)

    2018 arXiv

  57. [65]

    & Saenko, K

    Tan, S., Peng, X. & Saenko, K. Class-imbalanced Domain Adaptation: An Empirical Odyssey. Preprint at https://doi.org/10.48550/arXiv.1910.10320 (2020)

    2020 doi arXiv

  58. [66]

    Stokes, H. T. & Hatch, D. M. FINDSYM : program for identifying the space-group symmetry of a crystal. J Appl Crystallogr 38 , 237–238 (2005)

    2005

  59. [67]

    & Vecchio, K

    Kaufmann, K. & Vecchio, K. S. An Acquisition Parameter Study for Machine-Learning-Enabled Electron Backscatter Diffraction. Microsc Microanal 27 , 776–793 (2021)

    2021

  60. [68]

    L., Larsen, P

    Pang, E. L., Larsen, P. M. & Schuh, C. A. Global optimization for accurate determination of EBSD pattern centers. Ultramicroscopy 209 , 112876 (2020)

    2020

  61. [69]

    Deng, J. et al. ImageNet: A large-scale hierarchical image database. in 2009 IEEE Conference on Computer Vision and Pattern Recognition 248–255 (2009). doi:10.1109/CVPR.2009.5206848

    2009

  62. [70]

    & Loy, C

    Zhou, K., Liu, Z., Qiao, Y., Xiang, T. & Loy, C. C. Domain Generalization: A Survey. IEEE Trans. Pattern Anal. Mach. Intell. 1–20 (2022) doi:10.1109/TPAMI.2022.3195549

    2022

  63. [71]

    & Xiang, T

    Zhou, K., Yang, Y., Qiao, Y. & Xiang, T. Domain Adaptive Ensemble Learning. IEEE Trans. on Image Process. 30 , 8008–8018 (2021)

    2021

  64. [72]

    Debye-Waller factors

    Marquardt, K. et al. Quantitative electron backscatter diffraction (EBSD) data analyses using the dictionary indexing (DI) approach: Overcoming indexing difficulties on geological materials. American Mineralogist 102 , 1843–1855 (2017). Supplementary Material Analysis of resul...

    2017

  65. [73]

    Crystal symmetry determination in electron diffraction using machine learning,

    K. Kaufmann et al. , “Crystal symmetry determination in electron diffraction using machine learning,” Science , vol. 367, no. 6477, pp. 564–568, Jan. 2020, doi: 10.1126/science.aay3062. [2] A. Jain et al. , “Commentary: The Materials Project: A materials genome approach to acc...

    2020 doi

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.