Related Experiment Video
Updated: Aug 6, 2026
12:49
A Method for 3D Reconstruction and Virtual Reality Analysis of Glial and Neuronal Cells
Published on: September 28, 2019
Event-sparse stack denoising for 4D-STEM applications
Gregory Nordahl1, Rebekka Klemmt1, Martin Wibrand Larsen1
1Center for Sustainable Energy Materials (CENSEMAT), Aarhus University, DK-8000 Aarhus C, Denmark; Interdisciplinary Nanoscience Centre, Aarhus University, DK-8000 Aarhus C, Denmark.
Ultramicroscopy
|July 17, 2026
Summary
We developed a new denoising method for four-dimensional scanning transmission electron microscopy (4D-STEM) that improves data quality and reduces radiation exposure. This technique enhances defect detection sensitivity and material degradation analysis.
Area of Science:
- Materials Science
- Electron Microscopy
- Data Science
Background:
- Four-dimensional scanning transmission electron microscopy (4D-STEM) is crucial for nanoscale materials analysis.
- Conventional 4D-STEM data collection can be limited by noise and radiation dose.
- Bragg diffraction analysis in 4D-STEM requires high signal-to-noise ratio (SNR) data.
Purpose of the Study:
- To introduce a novel denoising method for 4D-STEM data, enhancing Bragg diffraction analysis.
- To reduce the radiation dose required for high-quality 4D-STEM data acquisition.
- To improve the sensitivity of defect detection and enable material degradation studies.
Main Methods:
- Developed event-sparse stack denoising for 4D-STEM, creating five-dimensional locally time-resolved STEM (LTR-STEM) datasets.
- Applied two sparsity denoising pipelines: DBSCAN with persistence thresholding and sparse PCA with single-step thresholding.
- Evaluated denoising performance using simulated SNR, recall curves, virtual imaging, and defect detection sensitivity.
Main Results:
- Sparsity-denoised 4D-STEM data achieved comparable SNR to raw data at only 38.5% of the exposure time, indicating lower dose requirements.
- A 4.1x increase in sensitivity to relative radial disk shift was observed for defect detection using denoised data.
- The LTR-STEM technique demonstrated potential for inspecting material degradation and estimating critical doses.
Conclusions:
- Event-sparse stack denoising effectively reduces noise in 4D-STEM data, enabling lower dose imaging.
- The LTR-STEM approach enhances defect detection capabilities and provides new avenues for material degradation analysis.
- This method offers a significant advancement for high-resolution structural analysis in materials science.
