AI for Transmission Electron Microscopy

Transmission electron microscopy (TEM) is an indispensable tool in modern materials science, life sciences, and physics, offering unparalleled resolution and versatile analytical capabilities for structural analysis. The pursuit of the next generation of advanced TEM capable of high throughput, atomic-scale quantification of materials and biological systems in their native states presents unprecedented challenges, such as the complexity of microscope operation which relies heavily on expert knowledge, the low experimental throughput due to manual operation, and the immense analytical burden posed by massive, high-dimensional datasets. Concurrently, artificial intelligence (AI), particularly deep learning, has made breakthrough progress in many areas. Its powerful capabilities in complex pattern recognition, big data analysis, and automated decision-making offer a revolutionary approach to overcome these bottlenecks in TEM. In this review, we summarize the recent advancements in the integration of TEM with AI, present our perspective on the current framework and propose a future outlook for AI–TEM.