Fig. 2: Schematic of 4D-STEM experiment and data analysis workflow. | npj Computational Materials

Fig. 2: Schematic of 4D-STEM experiment and data analysis workflow.

From: Unsupervised machine learning and cepstral analysis with 4D-STEM for characterizing complex microstructures of metallic alloys

Fig. 2

At each electron beam scan position, a nanobeam electron diffraction (NBED) pattern is acquired by the EMPAD and stored in a 4D dataset. PCA or NMF with K-means clustering is applied to the NBED dataset to evaluate phase segregation capability. A cepstral transform is then applied to the NBED dataset and run through the same clustering algorithm to compare the performance of all four combinations.

Back to article page