Automatic segmentation of STEM images/videos

Learning more

If you want to learn more about this project, feel free to check out our manuscript, our presentation for BiGmax workshop 2020 and our code written in python and c/c++.

Introduction

We present an unsupervised machine learning approach for segmentation of atomic-resolution microscopy images and videos. We combine symmetry-based local descriptors with classical unsupervised machine learning algorithms and highlight that the microscopy images might be segmented in an unsupervised manner. We demonstrate in this paper the successful application to the high-angle annular dark-field scanning Transmission electron microscopy (HAADF-STEM) images with atomic resolution. We release our code as a python module that reads the microscopy images and outputs the labels for all pixels, which is flexible to be used either as a standalone python code or as a plugin to other microscopy packages.