# Partial Scanning Transmission Electron Microscopy with Deep Learning

## TL;DR

This paper introduces a deep learning approach using a multiscale generative adversarial network to reconstruct full electron micrographs from partial scans, significantly reducing scan time and electron exposure while maintaining image quality.

## Contribution

Developed a novel two-stage multiscale GAN for partial STEM image reconstruction, trained on a large new dataset, enabling substantial reduction in scan time with minimal loss of detail.

## Key findings

- Achieved up to 87-fold reduction in scan coverage with only 6.2% error.
- Reduced scan time by 17.9 times with 3.8% error using spiral scans.
- Provided publicly available dataset, code, and models for further research.

## Abstract

Compressed sensing algorithms are used to decrease electron microscope scan time and electron beam exposure with minimal information loss. Following successful applications of deep learning to compressed sensing, we have developed a two-stage multiscale generative adversarial neural network to complete realistic 512$\times$512 scanning transmission electron micrographs from spiral, jittered gridlike, and other partial scans. For spiral scans and mean squared error based pre-training, this enables electron beam coverage to be decreased by 17.9$\times$ with a 3.8\% test set root mean squared intensity error, and by 87.0$\times$ with a 6.2\% error. Our generator networks are trained on partial scans created from a new dataset of 16227 scanning transmission electron micrographs. High performance is achieved with adaptive learning rate clipping of loss spikes and an auxiliary trainer network. Our source code, new dataset, and pre-trained models have been made publicly available at https://github.com/Jeffrey-Ede/partial-STEM