Engineering and Technology

A convolutional neural network for defect classification in Bragg coherent X-ray diffraction

B. Lim, E. Bellec, et al.

This paper highlights a groundbreaking 3D convolutional neural network (CNN) designed for swift and precise identification of defects in nanocrystals by analyzing Bragg coherent X-ray diffraction patterns. With training on extensive simulations, the CNN adeptly identifies dislocation types, marking a significant leap towards automated defect detection in materials science. This innovative research was conducted by Bruce Lim, Ewen Bellec, Maxime Dupraz, Steven Leake, Andrea Resta, Alessandro Coati, Michael Sprung, Ehud Almog, Eugen Rabkin, Tobias Schülli, and Marie-Ingrid Richard.