Generative adversarial networks for creating realistic training data for machine learning-based segmentation of FIB tomography data

Accurate 3D reconstruction of the structure of nanomaterials is essential for studying their physical properties. Focused ion beam (FIB) tomography is a preferred method for creating 3D image stacks of micrometer-sized material volumes at nanometer resolution. To achieve valid 3D reconstructions from FIB tomography data, semantic segmentation of these images using machine learning-based methods is often beneficial. However, supervised machine learning requires a large amount of training data and ground truth, which is challenging because FIB tomography is a destructive technique. This motivates the use of synthetic training data generated with Monte Carlo simulations of the FIB tomography process. However, these simulations are computationally expensive, and the resulting synthetic imaging data still differs from real FIB tomography data in terms of the statistical distribution of various features. In this study, we propose a novel approach to overcome both problems, i.e. requiring high computation time and difference in data distribution, using generative adversarial networks.

Subjects

3D reconstruction | domain adaptation | fast simulation | FIB-SEM tomography | synthetic data

DDC Class

620.1: Engineering Mechanics and Materials Science

006: Special computer methods

Funding(s)

SFB 986: Teilprojekt B09 - Mikrostrukturbasierte Klassifizierung und elektronenmikroskopische Analyse nanoporöser Metalle durch maschinelles Lernen

SFB 986: Zentralprojekt Z03 - Elektronenmikroskopie an multiskaligen Materialsystemen

Publication version

publishedVersion

Lizenz

https://creativecommons.org/licenses/by/4.0/

Name

Sardhara_2025_Mach._Learn.__Sci._Technol._6_025023.pdf

Type

Main Article

Size

3.63 MB

Format

Adobe PDF

Options

Generative adversarial networks for creating realistic training data for machine learning-based segmentation of FIB tomography data