DataVXMicroscopy and MicroanalysisPaper
journal article·2024-04-29Open access

Machine Learning Data Augmentation Strategy for Electron Energy Loss Spectroscopy: Generative Adversarial Networks

Daniel del-Pozo-Bueno, Demie Kepaptsoglou, Quentin M Ramasse, Francesca Peiró, Sònia Estradé

Microscopy and Microanalysis· Volume 30· Issue 2
Publisher pageOpen full text
Source abstract

Abstract

Abstract Recent advances in machine learning (ML) have highlighted a novel challenge concerning the quality and quantity of data required to effectively train algorithms in supervised ML procedures. This article introduces a data augmentation (DA) strategy for electron energy loss spectroscopy (EELS) data, employing generative adversarial networks (GANs). We present an innovative approach, called the data augmentation generative adversarial network (DAG), which facilitates data generation from a very limited number of spectra, around 100. Throughout this study, we explore the optimal configuration for GANs to produce realistic spectra. Notably, our DAG generates realistic spectra, and the spectra produced by the generator are successfully used in real-world applications to train classifiers based on artificial neural networks (ANNs) and support vector machines (SVMs) that have been successful in classifying experimental EEL spectra.
Source: crossref · Quality: baseline

Research context

OpenAlex topics: Machine Learning in Materials Science · Electron and X-Ray Spectroscopy Techniques · Machine Learning and Data Classification

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

Affiliations

LENS-MIND, Departament d'Enginyeria Electrònica i Biomèdica, Universitat de Barcelona , 1-11 Martí i Franquès, 08028 Barcelona , Spain

Institute of Nanoscience and Nanotechnology (IN2UB), Universitat de Barcelona , 1-11 Martí i Franquès, 08028 Barcelona , Spain

SuperSTEM Laboratory, Sci-Tech Daresbury Campus , Keckwick Lane, Daresbury WA4 4AD , UK

School of Physics, Engineering and Technology, University of York , Newton way, YO10 5DD Heslington , UK

Schools of Chemical and Process Engineering & Physics and Astronomy, Woodhouse Lane, University of Leeds , LS2 9JT Leeds , UK

Funding

Spanish Project · European Union NextGenerationEU/PRTR · MICIIN · ELECMI—ICTS Electron Microscopy for Materials Science · Generalitat de Catalunya · AGAUR agency of the Generalitat de Catalunya · ICREA Academia 2022 grant · UK National Research Facility for Advanced Electron Microscopy · Engineering and Physical Sciences Research Council

References (34)

Show source-provided references
1. Arora · 2017
2. Strategies for EELS data analysis. Introducing UMAP and HDBSCAN for dimensionality reduction and clustering10.1017/S1431927621013696 ↗
3. Extracting information from sequences of spatially resolved EELS spectra using multivariate statistical analysis10.1016/S0304-3991(99)00042-X ↗
4. Independent component analysis: A new possibility for analysing series of electron energy loss spectra10.1016/j.ultramic.2004.11.003 ↗
5. Bowles · 2018
6. Brock · 2018
7. Towards calibration-invariant spectroscopy using deep learning10.1038/s41598-019-38482-1 ↗
8. Electron-energy-loss-spectroscopy near-edge fine structures in the iron-oxygen system10.1103/PhysRevB.44.11402 ↗
9. Comparative of machine learning classification strategies for electron energy loss spectroscopy: Support vector machines and artificial neural networks10.1016/j.ultramic.2023.113828 ↗
10. Support vector machine for EELS oxidation state determination10.1016/j.ultramic.2020.113190 ↗
11. Direct evidence of a graded magnetic interface in bimagnetic core/shell nanoparticles using electron magnetic circular dichroism (EMCD)10.1021/acs.nanolett.1c02089 ↗
12. Mapping titanium and tin oxide phases using EELS: An application of independent component analysis10.1016/j.ultramic.2010.10.001 ↗
13. de la Peña · 2022
14. GAN-based synthetic medical image augmentation for increased CNN performance in liver lesion classification10.1016/j.neucom.2018.09.013 ↗
15. Goodfellow · 2014
16. The unreasonable effectiveness of data10.1109/MIS.2009.36 ↗
17. Heusel · 2017
18. Advanced elastic and reservoir properties prediction through generative adversarial network
19. α-EGAN: α-energy distance GAN with an early stopping rule10.1016/j.cviu.2023.103748 ↗
20. Kingma · 2014
21. A review: Data pre-processing and data augmentation techniques10.1016/j.gltp.2022.04.020 ↗
22. Precise size control of the growth of Fe 3 O 4 nanocubes over a wide size range using a rationally designed one-pot synthesis10.1021/acsnano.9b01281 ↗
23. Spin canting across core/shell Fe3O4/MnxFe3−xO4 nanoparticles10.1038/s41598-018-21626-0 ↗
24. RapidEELS: Machine learning for denoising and classification in rapid acquisition electron energy loss spectroscopy10.1038/s41598-021-97668-8 ↗
25. Toward laser-induced tuning of plasmonic response in high aspect ratio gold nanostructures10.1515/nanoph-2022-0193 ↗
26. Salimans · 2016
27. Data augmentation using generative adversarial networks (CycleGAN) to improve generalizability in CT segmentation tasks10.1038/s41598-019-52737-x ↗
28. Szegedy · 2015
29. Oxidation state and chemical shift investigation in transition metal oxides by EELS10.1016/j.ultramic.2012.03.002 ↗
30. Tanaka · 2019
31. 3D visualization of the iron oxidation state in FeO/Fe3O4 core–shell nanocubes from electron energy loss tomography10.1021/acs.nanolett.6b01922 ↗
32. Clustering analysis strategies for electron energy loss spectroscopy (EELS)10.1016/j.ultramic.2017.11.010 ↗
33. On data augmentation for GAN training10.1109/TIP.2021.3049346 ↗
34. EEL spectroscopic tomography: Towards a new dimension in nanomaterials analysis10.1016/j.ultramic.2012.07.020 ↗