4D STEM data analysis with KMeans clustering

Abstract

When using 4D STEM methods to study various material characteristics, large amounts of diffraction images are created for each sample studied. To determine different characteristics of the material locally from the data obtained from the diffraction patterns, it has been considered to use clustering machine learning algorithms that will be able to quickly read and classify all diffraction images. A KMeans algorithm has been adapted to classify this type of data. The method has been found to work satisfactorily when applied to an experimental example