Work

Accelerating domain-aware electron microscopy analysis using deep learning models with synthetic data and image-wide confidence scoring

Year: 2025

Type: article

Abstract: The integration of machine learning (ML) models enhances the efficiency, affordability, and reliability of feature detection in microscopy, yet their development and applicability are hindered by the dependency on scarce and often flawed manually labeled datasets with a lack of domain awareness. We addressed these challenges by creating a physics-based synthetic image and data generator, resulting in an ML model that achieves comparable precision (0.86), recall (0.63), F1 scores (0.71), and engineering property predictions ( R 2 = 0.82) to a model trained on human-labeled data. We enhanced both models by using feature prediction confidence scores to derive an image-wide confidence metric, enabling simple thresholding to eliminate ambiguous and out-of-domain images, resulting in performance boosts of 5–30% with a filtering-out rate of 25%. Our study demonstrates that synthetic data can eliminate human reliance in ML and provides a means for domain awareness in cases where many feature detections per image are needed.

Source: npj Computational Materials

DOI: 10.1038/s41524-025-01756-6

Authors: M. J. Lynch, R. Jacobs, G. A. Bruno, P. Patki, D. Morgan, K. G. Field

Institutions: University of Michigan, University of Wisconsin–Madison, Intel (United States)

Language: English

Topic: Advanced Electron Microscopy Techniques and Applications

Subfield: Structural Biology

Field: Biochemistry, Genetics and Molecular Biology

Domain: Life Sciences

Open Access status: gold

Estimated APC paid: $3,490

Funders: National Aeronautics and Space Administration, Electric Power Research Institute, Nuclear Safety and Security Commission