Remote PACBED Thickness Determination by CNNs

A convenient thickness determination technique for crystalline samples is the position averaged convergent-beam electron diffraction(PACBED) method [1]. The thickness is determined by finding the best match of the recorded PACBED pattern with a series ofsimulated PACBEDs. This process can be automatized by convolutional neural networks (CNNs), making the method fast and easy toapply [2]. However, the simulation of a synthetic dataset and the training of the CNNs have high computational cost and these CNNsare only valid for the specific trained system. Therefore, we built a working prototype of a server-based thickness determination byCNNs with a shared CNN-database and a GUI. By this, every scientist, even without knowledge about machine learning andmultislice simulation, could determine the specimen thickness by PACBEDs within few seconds during a microscope session.