Defect detection in atomic-resolution images via unsupervised learning with translational invariance

Published

November 15, 2021

Author(s)

Yueming Guo, Sergei Kalinin, Cai Hui, Kai Xiao, Sergiy Krylyuk, Albert Davydov, Qianying Guo, Andrew Lupini

Abstract

Crystallographic defects can now be routinely imaged at atomic resolution with aberration-corrected scanning transmission electron microscopy (STEM) at high speed, with the potential for vast volumes of data to be acquired in relatively short times, or through autonomous experiments that can continue over very long periods. Therefore, automatic detection and classification of defects in the images is now needed in order to handle the data in an efficient way. However, just like many other tasks of object detection and identification in artificial intelligence, it is challenging to detect and identify defects from STEM images. Furthermore, it is difficult to deal with crystal structures that have many atoms and low symmetries. Previous methods for defect detection and classification were based on supervised learning, which requires human-labeled data. In this work, we develop an approach for defect detection with unsupervised machine learning based on a one-class support vector machine (OCSVM). We introduce two schemes of image segmentation and data preprocessing, which both involve taking the Patterson function of each segment as inputs. We demonstrate that this method can be applied to various defects, such as point and line defects in two-dimensional (2D) materials and twin boundaries in three-dimensional (3D) nanocrystals.

Citation

npj Computational Materials

Volume

7

Pub Type

Journals

Download Paper

https://doi.org/10.1038/s41524-021-00642-1

Local Download

Keywords

Machine learning, Crystallographic Defects, STEM

Citation

Guo, Y. , Kalinin, S. , Hui, C. , Xiao, K. , Krylyuk, S. , Davydov, A. , Guo, Q. and Lupini, A. (2021), Defect detection in atomic-resolution images via unsupervised learning with translational invariance, npj Computational Materials, [online], https://doi.org/10.1038/s41524-021-00642-1, https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=932610 (Accessed October 5, 2025)