Skip to main content
Fig. 3 | Advanced Structural and Chemical Imaging

Fig. 3

From: Identifying local structural states in atomic imaging by computer vision

Fig. 3

Classification of Local Structural States. Detected atomic columns at different noise levels in simulated STEM images are classified by hierarchical clustering, with different structural classes represented by circles with different colors. The different atomic columns in the [100] projection of the SrTiO3 unit cell are all classified as distinct structural states by the presented approach (a). In the presence of noise, the distinction between Sr and Ti atomic columns is still maintained (b). Note that Sr atoms at the edge of the image belong to separate classes since their coordination is different than that of Sr atoms in the “bulk”. c Classification of atoms in the image of a SrTiO3/BaTiO3 interface distinguishes the interfacial atoms (Sr, Ti, Ba) from those present in the bulk phases, and provides a complete description of the structural configurations present in the image

Back to article page