Pith. sign in

REVIEW 4 major objections 5 minor 42 references

Novel Concept-Oriented Synthetic Data approach for Training Generative AI-Driven Crystal Grain Analysis Using Diffusion Model

T0 review · 4 major / 5 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 A synthetic-data pipeline trains a diffusion model to map crystal grains from low-resolution TEM images at 97.23% accuracy.

desk verdict The reported 97.23% accuracy is not yet verifiable because Eq. (1) leaves boundary thickness and manual marking unspecified; the paper's real contribution is the clearly specified 'concept-oriented' synthetic data recipe, which deserves a serious look with a stricter evaluation. read the letter →

arxiv 2504.14782 v1 pith:6HDXP7OU submitted 2025-04-21 cs.LG cond-mat.mtrl-sci

Ahmed Sobhi Saleh , Kristof Croes , Hajdin Ceric , Ingrid De Wolf , Houman Zahedmanesh This is my paper · ORCID
classification cs.LGcond-mat.mtrl-sci
keywords MachineLearningGenerativeAIDiffusionModelSyntheticDataGrainBoundaryDetectionTransmissionElectronMicroscopyMicrostructureAnalysisVoronoiTessellation
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 presents a fully automated pipeline for extracting polycrystalline grain boundaries from low-resolution TEM images, replacing the manual or high-end experimental methods normally used for crystal orientation mapping. To overcome the shortage of labelled microscopy data, the authors generate 3,000 synthetic TEM images from a Monte Carlo and Voronoi-based simulator that encodes the physical concepts a grain-boundary detector needs. The pipeline combines three classical edge detectors, an encoder-decoder network that proposes an initial boundary map, and a diffusion model that iteratively removes noise and reconnects broken boundaries. On benchmark molybdenum images, the extracted grain morphologies match experimentally obtained EBSD/IPF maps with an average accuracy of 97.23%, outperforming a reported 89% for a CNN-based software approach. If the result transfers, it would make high-throughput grain analysis possible from routine microscopy images.

What carries the argument

The load-bearing object is the concept-oriented synthetic data generator, a seven-stage recipe applied to TEM images. It starts from a lognormal grain-size distribution, packs circles with a Monte Carlo dynamic-packing algorithm, and converts the circle centres and radii into a weighted Voronoi (power) diagram whose cells are convex polygons, each receiving a random grey level; a faint flipped copy of the image is added to simulate depth-overlap artefacts, followed by salt-and-pepper, Gaussian, and Poisson noise and a median filter. This yields paired training data: a binary ground-truth boundary map and a synthetic TEM-like input. The inference pipeline then runs Sobel, Laplacian-of-Gaussian, and Canny edge detectors, feeds their concatenated RGB map into an encoder-decoder network (Net1) that produces an initial binary guess, and passes that guess to a second encoder-decoder diffusion network (Net2) that iteratively denoises and reconnects boundaries until a convergence threshold is met.

What would settle it

Take the trained pipeline, without retraining, to a TEM image of a microstructure containing elongated, twinned, or non-convex grains whose true boundaries are known from EBSD/IPF, and compute Eq. (1). If the detected-boundary accuracy falls substantially below 97.23%, the convex-polygon concept in the synthetic training data is not sufficient for real microstructures.

Watch

Extended reading notes

Core claim

The central claim is that a diffusion model trained exclusively on concept-oriented synthetic data can turn a noisy low-resolution TEM image into a complete grain-boundary map whose accuracy is comparable to that of advanced experimental characterization. The claim is quantified on three molybdenum cases with ground truth from EBSD inverse pole figure maps validated by XRD: the detected boundary length matches 97.2%, 97.1%, and 97.4% of the true boundaries, giving a 97.23% average. The authors further argue that the same seven-stage synthetic-data recipe generalizes: encode the concepts a model must know, generate targets with a Monte Carlo simulator, render them into input-like examples, and add controlled noise.

Load-bearing premise

The load-bearing assumption is that the few encoded concepts—convex polygon grains, boundaries visible as brightness differences, faint overlap edges to be ignored, and the three noise types—are enough to represent real TEM images, so a network trained only on synthetic images will recognize real grain boundaries.

Editorial extensions

If this is right

  • Grain-boundary maps for metals could be produced automatically from routine TEM images, without the sample preparation and instrumentation needed for EBSD, HRTEM, or PED.
  • The diffusion model is used as a structure predictor that refines an initial guess, not as a generator from pure noise, so the same iterative repair strategy can be applied to other segmentation tasks with scarce data.
  • The seven-stage synthetic-data recipe gives a concrete template for encoding domain concepts into training examples whenever labelled real images are insufficient.
  • Because the synthetic data controls noise and boundary appearance, models can be trained to ignore faint depth-overlap edges that human annotators would otherwise need to label manually.

Reading between the lines

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

  • The reported 97.23% is measured on molybdenum only; the transfer to other metals (Cu, Ru) is demonstrated qualitatively, so a quantitative accuracy study on those metals is a natural next check.
  • Because the synthetic grains are convex Voronoi polygons, the pipeline would likely need additional concepts to handle heavily twinned, abnormally grown, or otherwise non-convex grain structures; this is a testable limitation, not a claim in the paper.
  • The accuracy metric's boundary-thickness choice changes the score; a sensitivity analysis over boundary thickness would show how much of the 97.23% is definitional.
  • The concept-oriented recipe could be applied to other imaging modalities with known physical priors, such as SEM or optical microscopy, by adjusting the noise and concept-encoding stages.
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 / 5 minor

Summary. The paper proposes a three-stage pipeline for automated grain-boundary extraction from TEM images: three classical edge detectors (Sobel, LoG, Canny) whose outputs are stacked into an RGB image, an encoder-decoder network (Net1) that turns this RGB image into a binary initial guess, and a second encoder-decoder network (Net2) used iteratively as a diffusion-style refiner that denoises and connects the boundary map. To address data scarcity, the authors introduce a seven-stage 'concept-oriented' synthetic-data methodology and generate 3,000 synthetic TEM images via a Monte Carlo Voronoi procedure with prescribed grain-size distributions, brightness assignments, overlap effects, and noise models. The model is trained and tested on these synthetic images, then applied to real TEM images of Cu, Ru, and Mo. The quantitative claim is an average accuracy of 97.23% measured on three Mo images by comparing predicted boundary pixels against EBSD/IPF-derived ground truth from an external reference.

Significance. The core idea is attractive and timely: a fully synthetic training set for microstructure segmentation, combined with a diffusion-based iterative refinement stage, could be a practical answer to the scarcity of labeled microscopy data. The synthetic-data generation pipeline is described in enough detail to be reproduced, and the use of independent EBSD/IPF ground truth from ref. [29] is a genuine strength, since the central evaluation is not self-referential. The claim of 97.23% accuracy, if made reproducible with a fully specified metric and a broader benchmark, would be a useful contribution to automated microstructure analysis. As presented, however, the quantitative evidence is too thin and the accuracy metric too under-specified to support the abstract's headline number.

major comments (4)
  1. [Model benchmarking, Eq. (1)] The accuracy metric in Eq. (1) is not reproducible as stated. The metric is described as the percentage of correctly detected boundary pixels within a 'uniform grain boundary thickness', but the thickness value, the procedure for registering predicted boundaries to the IPF ground truth, and the rule for counting pixels in 'missed' boundaries are never specified. The manual orange markings in Fig. 3 are presumably used to define the denominator, but the paper does not say how those markings were created or validated. Because a larger tolerance band inflates the number of correctly matched pixels, the reported 97.23% average cannot be independently verified. Please specify the exact tolerance in pixels or physical units, the matching/registration algorithm, and the manual-marking protocol, or replace the metric with a standard segmentation metric such as F1 score at a fixed tolerance.
  2. [Model benchmarking and Table 1] The quantitative benchmark consists of only three Mo TEM images from a single external publication, and the Cu and Ru results in Fig. 2 are presented qualitatively without any accuracy measurement. The abstract's statement that the model was applied to 'various metals' with the reported accuracy is therefore not supported by the controlled evaluation. Please either add quantitative results for additional images and materials or explicitly restrict the accuracy claim to the three Mo examples.
  3. [Methods B, concepts (i)-(iii)] The synthetic training distribution relies on three strong assumptions: grains are convex polygons, boundaries appear mainly between regions of different brightness, and faint overlap edges should be ignored. These assumptions may not hold for real TEM images of other materials or for images with non-convex grains, thickness variations, or specimen preparation artifacts. No ablation or synthetic-to-real transfer experiment is provided to demonstrate that the pipeline remains accurate when such assumptions are violated. This is load-bearing for the claimed generality of the method; please validate on a more diverse set of real images with quantitative metrics, or clearly state these assumptions as limitations in the conclusions.
  4. [Model benchmarking and Results] The three accuracy values in Table 1 (97.2%, 97.1%, 97.4%) are reported without any uncertainty or variability estimate, and no baseline method is evaluated on the same three images. The comparison to the 89% accuracy of UNet+CHAC is cross-study and not controlled, since it comes from a different dataset. Please add per-image repeated-run or bootstrapped error estimates and run at least one standard segmentation or grain-boundary baseline on the same Mo images so that the improvement claim is directly supported.
minor comments (5)
  1. [Methods C, Fig. 5] The text uses 'transverse convolution' where the standard term is 'transposed convolution'; please correct this terminology throughout.
  2. [Eq. (1) and surrounding text] The text says the metric measures the 'length' of detected grain boundaries, but Eq. (1) is defined in terms of pixel counts; please clarify the relationship between pixel count and boundary length, or align the wording with the equation.
  3. [Conclusion] There is a typo: 'This study provide s an advancement' should be 'This study provides an advancement'.
  4. [Fig. 2 caption] The phrase 'four equidistant images from a total of 7 iterations' would be clearer as 'four equally spaced images from a total of 7 iterations'.
  5. [References] The citation to ref. [2] for the lognormal grain-size distribution in Methods B seems out of place, since that reference is about electromigration modeling methods; consider citing a more direct source for grain-size distribution statistics.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the benchmark is anchored to external EBSD/IPF ground truth from ref [29], so the 97.23% claim is not self-referential; only a minor non-load-bearing self-citation appears in the synthetic-data motivation.

full rationale

The paper's central derivation chain is not circular. Synthetic TEM images in Methods B are generated from explicit physical concepts: Voronoi/convex grains, brightness-contrast boundaries, faint overlap edges, and specified noise models. These synthetic data are not fitted to the three test images, and the test images and ground-truth grain-boundary maps come from the external study ref [29]. The accuracy metric in Eq. (1) compares predicted boundary pixels with boundaries derived from EBSD/IPF maps, so the reported 97.23% is an external benchmark rather than a quantity defined by the model's own inputs. The only self-citation is ref [2], cited in Methods B for the lognormal grain-size distribution; it shares authors with the present paper, but it is accompanied by the external ref [36] and it does not enter Eq. (1) or the benchmarking procedure, so it is not load-bearing. Unspecified boundary thickness and manual marking of missed boundaries in Fig. 3 are reproducibility concerns, not circularity: no equation or construction in the paper reduces the predicted quantity to its own input. The claim about applicability to 'various metals' is broader than the three Mo cases actually benchmarked, but overclaiming is a correctness/caveat issue rather than circular reasoning.

Assumptions & free parameters 8 free parameters · 7 assumptions · 0 invented entities

The method depends on several hand-set noise and geometry parameters, plus domain assumptions about grain morphology and imaging physics. The lognormal distribution and convexity are supported by cited literature, but the exact parameter values and the sufficiency of the synthetic dataset are not externally validated. No new physical entities are introduced.

free parameters (8)
  • Salt-and-pepper noise density = 0.05 (5% of pixels)
    Hand-picked in Methods B to simulate TEM noise; not fitted to data.
  • Gaussian noise variance = 0.01
    Hand-picked in Methods B to simulate TEM noise.
  • Poisson noise parameter = mean set equal to original pixel value
    Poisson noise chosen in Methods B; not fitted.
  • Overlap weakening factor = 0.1
    Chosen in Methods B to model faint grain overlap in TEM.
  • Median filter kernel size = 5x5
    Smoothing kernel chosen in Methods B.
  • Grain boundary thickness in synthetic targets = not specified in text
    Boundary width used for binary target images is described but not quantified.
  • Convergence threshold of iterative refinement = 0.003 * total pixels * 255
    Threshold chosen via sensitivity analysis, but no analysis details given.
  • Lognormal grain size distribution parameters = mean and variance not specified
    Grain sizes drawn from lognormal distribution in step 4, but parameters are not reported.
assumptions (7)
  • standard math A weighted Voronoi tessellation (power diagram) of circle centers produces convex polygonal grains.
    Invoked in Methods B, step 5 to generate grain morphology.
  • domain assumption Grain size distributions in polycrystalline materials are lognormal.
    Stated in Methods B, step 4 with citations [2], [36]; used as the primary high-level feature.
  • domain assumption Grains in polycrystalline microstructures are convex polygons due to energy minimization.
    Concept (i) in Methods B, used to justify Voronoi generation.
  • domain assumption Grain boundaries generally appear as brightness changes in TEM images, and symmetric orientations can have similar brightness.
    Concept (ii) in Methods B, used to encode brightness variations in synthetic images.
  • domain assumption Faint boundaries from overlapping grains at different depths should be disregarded.
    Concept (iii) in Methods B, encoded via a flipped, weakened duplicate image.
  • ad hoc to paper The synthetic images capture the essential features of real TEM images for grain boundary detection.
    Core assumption of the concept-oriented approach; not independently validated for diverse real imaging conditions.
  • domain assumption EBSD/IPF maps from ref [29] provide accurate ground truth grain boundaries for the test TEM images.
    Used in Model benchmarking to compute Eq. (1) accuracy.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Novel Concept-Oriented Synthetic Data approach for Training Generative AI-Driven Crystal Grain Analysis Using Diffusion Model." pith.science (2026). https://pith.science/paper/6HDXP7OU

@misc{pith2026250414782,
  author       = {Pith},
  title        = {Pith review of: Novel Concept-Oriented Synthetic Data approach for Training Generative AI-Driven Crystal Grain Analysis Using Diffusion Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6HDXP7OU}},
  note         = {Machine review of arXiv:2504.14782}
}
read the original abstract

The traditional techniques for extracting polycrystalline grain structures from microscopy images, such as transmission electron microscopy (TEM) and scanning electron microscopy (SEM), are labour-intensive, subjective, and time-consuming, limiting their scalability for high-throughput analysis. In this study, we present an automated methodology integrating edge detection with generative diffusion models to effectively identify grains, eliminate noise, and connect broken segments in alignment with predicted grain boundaries. Due to the limited availability of adequate images preventing the training of deep machine learning models, a new seven-stage methodology is employed to generate synthetic TEM images for training. This concept-oriented synthetic data approach can be extended to any field of interest where the scarcity of data is a challenge. The presented model was applied to various metals with average grain sizes down to the nanoscale, producing grain morphologies from low-resolution TEM images that are comparable to those obtained from advanced and demanding experimental techniques with an average accuracy of 97.23%.

Figures

Figures reproduced from arXiv: 2504.14782 by the authors.

Figure 2
Figure 2. Microstructure extraction from TEM images. a) A step-by-step demonstration of the microstructure extraction process from a TEM image of a Ru line. Residual noise is annotated by blue circles and broken grain boundaries are annotated by yellow circles. The residual noise fragments don’t lie on a true grain boundary whereas broken grain boundaries do. The final stack of images shows four equidistant images from a tota… view at source ↗
Figure 5
Figure 5. Layers architecture for Net1 and Net2. (A) The detailed layered architecture for Net1. (B) The detailed layered architecture for Net2. Conv stands for convolution process. TConv stands for transverse convolution process. Max-Pool stand for maxpooling process to reduce dimensionality. The dimension for each layer input is given above each data stage in the format (No. of channels)@(image width in pixels)×(image lengt… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 39 canonical work pages

  1. [29]

    Numerical evaluation of grain boundary electron scattering in molybdenum thin films: A critical analysis for advanced interconnects,

    S. H. Joo and D. Choi, "Numerical evaluation of grain boundary electron scattering in molybdenum thin films: A critical analysis for advanced interconnects," Vacuum, vol. 222, p. 113025, 2024

    work page 2024

  2. [1]

    Clemens, S

    H. Clemens, S. Mayer and C. Scheu, Microstructure and Properties of Engineering Materials, Weinheim, Germany: Wiley-VCH Verlag GmbH & Co., 2017

    work page 2017

  3. [2]

    Review—Modeling Methods for Analysis of Electromigration Degradation in Nano-Interconnects,

    H. Ceric, S. Selberherr, H. Zahedmanesh, R. L. d. Orio and K. Croes, "Review—Modeling Methods for Analysis of Electromigration Degradation in Nano-Interconnects," ECS Journal of Solid State Science and Technology, vol. 10, no. 3, p. 035003, 2021

    work page 2021

  4. [3]

    Effect of metal line width on electromigration of BEOL Cu interconnects,

    S. Choi, C. Christiansen, L. Cao, J. Zhang, R. Flippi, T. Shen, K. B. Yeap, S. Ogden, H. Zhang, B. Fu and P. Justison, "Effect of metal line width on electromigration of BEOL Cu interconnects," in IEEE (IRPS), Burlingame, CA, 2018

    work page 2018

  5. [4]

    Microstructure Evolution and Effect on Resistivity for Cu Nanointerconnects and Beyond,

    S.-T. Hu, L. Cao, L. Spinella and P. S. Ho, "Microstructure Evolution and Effect on Resistivity for Cu Nanointerconnects and Beyond," in IEEE International Electron Devices Meeting (IEDM), San Francisco, CA, USA, 2018

    work page 2018

  6. [5]

    J. L. McCall, Scanning Electron Microscopy for Microstructural Analysis, Boston, MA: Springer, 1973

    work page 1973

  7. [6]

    J. S. Lim, Two-Dimensional Signal and Image Processing, Englewood Cliffs, NJ, Prentice Hall, 1990

    work page 1990

  8. [7]

    J. R. Parker, Algorithms for Image Processing and Computer Vision, New York: John Wiley & Sons, Inc., 1997

    work page 1997

Show all 42 references
  1. [8]

    Machine vision for three - dimensional scenes

    P. Danielsson and O. Seger, Generalized and Separable Sobel Operators,in "Machine vision for three - dimensional scenes",Herbert Freeman (ed), Academic Press, 1990

    1990

  2. [9]

    A Computational Approach to Edge Detection,

    J. Canny, "A Computational Approach to Edge Detection," IEEE Transactions on Pattern Analysis and Machine Intelligence, Vols. PAMI-8, no. 6, pp. 679-698, 1986

    1986

  3. [10]

    Automatic detection of particle size distribution by image analysis based on local adaptive canny edge detection and modified circular Hough transform,

    Y. Meng, Z. Zhang, H. Yin and T. Ma, "Automatic detection of particle size distribution by image analysis based on local adaptive canny edge detection and modified circular Hough transform," Micron, vol. 106, pp. 34-41, 2018

    2018

  4. [11]

    An automated methodology for grain segmentation and grain size measurement from optical micrographs,

    S. Banerjee, P. Chakraborti and S. Saha, "An automated methodology for grain segmentation and grain size measurement from optical micrographs," Measurement, vol. 140, no. 0263-2241, pp. 142-150, 2019

    2019

  5. [12]

    De-noising Filters for TEM (Transmission Electron Microscopy) Image of Nanomaterials,

    H. S. Kushwaha, S. Tanwar, K. Rathore and S. Srivastava, "De-noising Filters for TEM (Transmission Electron Microscopy) Image of Nanomaterials," in Second International Conference on Advanced Computing & Communication Technologies, 2012

    2012

  6. [13]

    New methods for automatic quantification of microstructural features using digital image processing,

    A. Campbell, P. Murray, E. Yakushina, S. Marshall and W. Ion, "New methods for automatic quantification of microstructural features using digital image processing," Materials & Design, vol. 141, no. 0264-1275, pp. 395-406, 2018

    2018

  7. [14]

    Deep learning object detection in materials science: Current state and future directions,

    R. Jacobs, "Deep learning object detection in materials science: Current state and future directions," Computational Materials Science, vol. 211, no. 0927-0256, p. 111527, 2022

    2022

  8. [15]

    Automated analysis of grain morphology in TEM images using convolutional neural network with CHAC algorithm,

    X. Xu, Z. Yu, W.-Y. Chen, A. Chen, A. Motta and X. Wang, "Automated analysis of grain morphology in TEM images using convolutional neural network with CHAC algorithm," Journal of Nuclear Materials, vol. 588, p. 154813, 2024

    2024

  9. [16]

    Deep unsupervised learning using nonequilibrium thermodynamics,

    J. Sohl-Dickstein, E. A. Weiss, N. Maheswaranathan and S. Ganguli, "Deep unsupervised learning using nonequilibrium thermodynamics," in 32nd International Conference on International Conference on Machine Learning, Lille, France, 2015

    2015

  10. [17]

    Denoising Diffusion Probabilistic Models,

    J. Ho, A. Jain and P. Abbeel, "Denoising Diffusion Probabilistic Models," ArXiv, vol. 2006.11239, 2020

    2006 arXiv

  11. [18]

    SDXK: improving latent diffusion models for high-resolution image synthesis,

    D. Podell, Z. English, K. Lacey, A. Blattmann, T. Dockhorn, J. Müller, J. Penna and R. Rombach, "SDXK: improving latent diffusion models for high-resolution image synthesis," in Twelfth International Conference on Learning Representations (ICLR), 2024

    2024

  12. [19]

    DiffWave: A Versatile Diffusion Model for Audio Synthesis,

    Z. Kong, W. Ping, J. Huang, K. Zhao and B. Catanzaro, "DiffWave: A Versatile Diffusion Model for Audio Synthesis," in International Conference on Learning Representations (ICLR), Vienna, Austria, 2021

    2021

  13. [20]

    Predicting sample size required for classification performance,

    R. L. Figueroa, Q. Zeng-Treitler, S. Kandula and L. H. Ngo, "Predicting sample size required for classification performance," BMC Medical Informatics and Decision Making, vol. 12, no. 1, p. 8, 2012

    2012

  14. [21]

    Chakravorti, Electric Field Analysis, CRC Press, 2017

    S. Chakravorti, Electric Field Analysis, CRC Press, 2017

    2017

  15. [22]

    An electrostatic study of curvature effects on electric field stress in high voltage differentials,

    N. Ma, V. F. Jones, E. J. Miller and K. J. Kozaczek, "An electrostatic study of curvature effects on electric field stress in high voltage differentials," Journal of Electrostatics, vol. 101, p. 103370, 2019

    2019

  16. [23]

    Kleppner and R

    D. Kleppner and R. Kolenkow, An Introduction to Mechanics (2nd Edition), Cambridge, UK: Cambridge University Press, 2013

    2013

  17. [24]

    Computer Simulation of Powder Compaction of Spherical Particles,

    D. He and N. Ekere, "Computer Simulation of Powder Compaction of Spherical Particles," Journal of Materials Science Letters, vol. 17, no. 20, pp. 1723-1725, 1998

    1998

  18. [25]

    Simulation of polycrystalline structure with Voronoi diagram in Laguerre geometry based on random closed packing of spheres,

    Z. Fan, Y. Wu, X. Zhao and Y. Lu, "Simulation of polycrystalline structure with Voronoi diagram in Laguerre geometry based on random closed packing of spheres," Computational Materials Science, vol. 29, no. 3, pp. 301-308, 2004

    2004

  19. [26]

    S. E. Umbaugh, Digital Image Processing and Analysis: Human and Computer Vision Applications with CVIPtools, Second Edition (2nd ed.), Boca Raton: CRC Press, 2010

    2010

  20. [27]

    An Experimental Analysis on the Sensitivity of the Most Widely Used Edge Detection Methods to Different Noise Types,

    M. N. Raihan, N. Ulfat and N. Saqib, "An Experimental Analysis on the Sensitivity of the Most Widely Used Edge Detection Methods to Different Noise Types," in 2nd International Conference on Computing Advancements (ICCA), 2022

    2022

  21. [28]

    A Review of Classic Edge Detectors,

    H. Spontón and J. Cardelino, "A Review of Classic Edge Detectors," Image Processing On Line, vol. 5, pp. 90-123, 2015

    2015

  22. [30]

    Consistency Models,

    Y. Song, P. Dhariwal, M. Chen and I. Sutskever, "Consistency Models," in International Conference on Machine Learning (ICML), Honolulu, Hawaii, 2023

    2023

  23. [31]

    Generative Modeling by Estimating Gradients of the Data Distribution,

    Y. Song and S. Ermon, "Generative Modeling by Estimating Gradients of the Data Distribution," in Conference on Neural Information Processing Systems (NeurIPS), Vancouver, Canada, 2019

    2019

  24. [32]

    Score-Based Generative Modeling through Stochastic Differential Equations,

    Y. Song, J. Sohl-Dickstein, D. P. Kingma, A. Kumar, S. Ermon and B. Poole, "Score-Based Generative Modeling through Stochastic Differential Equations," in International Conference on Learning Representations (ICLR), 2021

    2021

  25. [33]

    Two‐Dimensional Motion of Idealized Grain Boundaries,

    W. W. Mullins, "Two‐Dimensional Motion of Idealized Grain Boundaries," Journal of Applied Physics, vol. 27, no. 8, pp. 900-904, 1956

    1956

  26. [34]

    Grain Shapes and Other Metallurgical Applications of Topology,

    C. S. Smith, "Grain Shapes and Other Metallurgical Applications of Topology," Metallography, Microstructure, and Analysis, vol. 4, no. 6, p. 543–567, 2015

    2015

  27. [35]

    D. B. Williams and C. B. Carter, Transmission Electron Microscopy: A Textbook for Materials Science, New York: Springer, 2009

    2009

  28. [36]

    A strategy for synthetic microstructure generation and crystal plasticity parameter calibration of fine-grain-structured dual-phase steel,

    W. Liu, J. Lian, N. Aravas and S. Münstermann, "A strategy for synthetic microstructure generation and crystal plasticity parameter calibration of fine-grain-structured dual-phase steel," International Journal of Plasticity, vol. 126, p. 102614, 2020

    2020

  29. [37]

    A novel point inclusion test for convex polygons based on Voronoi tessellations,

    R. S. Zengin and V. Sezer, "A novel point inclusion test for convex polygons based on Voronoi tessellations," Applied Mathematics and Computation, vol. 299, p. 126001, 2021

    2021

  30. [38]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, "Adam: A method for stochastic optimization," arXiv, vol. arXiv:1412.6980, 2014

    2014 arXiv

  31. [39]

    Understanding the Generalization Benefit of Normalization Layers: Sharpness Reduction,

    K. Lyu, Z. Li and S. Arora, "Understanding the Generalization Benefit of Normalization Layers: Sharpness Reduction," ArXiv, vol. abs/2206.07085, 2022

    2022 arXiv

  32. [40]

    An Overview of Overfitting and its Solutions,

    X. Ying, "An Overview of Overfitting and its Solutions," Journal of Physics: Conference Series (JPCS), vol. 1168, no. 2, p. 022022, 2019

    2019

  33. [41]

    Attention Is All You Need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser and I. Polosukhin, "Attention Is All You Need," in 31st Conference on Neural Information Processing Systems (NIPS), Long Beach, CA, USA, 2017

    2017

  34. [42]

    Physics -informed machine learning,

    G. E. Karniadakis, I. G. Kevrekidis, L. Lu, P. Perdikaris, S. Wang and L. Yang, "Physics -informed machine learning," Nature Reviews Physics, vol. 3, no. 6, pp. 422-440, 2021. Acknowledgements Acknowledgements The authors extend their sincere gratitude to Marleen van der Veen ...

    2021

Pith tools

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