Quantifying noise limitations of neural network segmentations in high-resolution transmission electron microscopy

Matthew Helmi Leth Larsen, William Bang Lomholdt, Cuauhtemoc Nuñez Valencia, Thomas W. Hansen, Jakob Schiøtz*

*Corresponding author for this work

Research output: Contribution to journalJournal articleResearchpeer-review

36 Downloads (Pure)

Abstract

Motivated by the need for low electron dose transmission electron microscopy imaging, we report the optimal frame dose (i.e. e2) range for object detection and segmentation tasks with neural networks. The MSD-net architecture shows promising abilities over the industry standard U-net architecture in generalising to frame doses below the range included in the training set, for both simulated and experimental images. It also presents a heightened ability to learn from lower dose images. The MSD-net displays mild visibility of a Au nanoparticle at 20–30 e2, and converges at 200 e2 where a full segmentation of the nanoparticle is achieved. Between 30 and 200 e2 object detection applications are still possible. This work also highlights the importance of modelling the modulation transfer function when training with simulated images for applications on images acquired with scintillator based detectors such as the Gatan Oneview camera. A parametric form of the modulation transfer function is applied with varying ranges of parameters, and the effects on low electron dose segmentation is presented.
Original languageEnglish
Article number113803
JournalUltramicroscopy
Volume253
Number of pages8
ISSN0304-3991
DOIs
Publication statusPublished - 2023

Keywords

  • Beam damage
  • HR-TEM
  • Machine learning
  • Modulation transfer function
  • Signal-to-noise

Fingerprint

Dive into the research topics of 'Quantifying noise limitations of neural network segmentations in high-resolution transmission electron microscopy'. Together they form a unique fingerprint.

Cite this