Atomic Depth Estimation From Noisy Electron Microscopy Data Via Deep Learning

2026-01-20

AI Summary

Recovering atomic-scale three-dimensional depth information from highly noisy transmission electron microscopy (TEM) images remains a formidable challenge. This work addresses this problem by, for the first time, formulating atomic depth estimation as a semantic segmentation task and proposing a deep convolutional neural network–based approach. Trained on simulated TEM data augmented with synthetic noise, the model generates pixel-wise depth segmentation maps to predict the depth of atomic columns in CeO₂ nanoparticles. The method achieves high accuracy, well-calibrated depth estimates, and strong noise robustness on both simulated and experimental TEM datasets, significantly enhancing the reliability and precision of atomic-scale depth reconstruction under high-noise conditions.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: State Estimation

Application Category

• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Security and Privacy: Data transparency and provenance

Abstract

We present a novel approach for extracting 3D atomic-level information from transmission electron microscopy (TEM) images affected by significant noise. The approach is based on formulating depth estimation as a semantic segmentation problem. We address the resulting segmentation problem by training a deep convolutional neural network to generate pixel-wise depth segmentation maps using simulated data corrupted by synthetic noise. The proposed method was applied to estimate the depth of atomic columns in CeO2 nanoparticles from simulated images and real-world TEM data. Our experiments show that the resulting depth estimates are accurate, calibrated and robust to noise.

Problem

• atomic depth estimation
• noisy electron microscopy
• 3D atomic information
• TEM image analysis
• depth segmentation

Innovation

• atomic depth estimation
• deep learning
• semantic segmentation
• noisy TEM data
• convolutional neural network