New Machine-Learning Framework Improves Real-Time Tracking of Nanoscale Defects in Materials
Researchers have developed a novel analysis framework that combines machine learning with particle filter estimation to quantitatively evaluate dynamic defect behavior from real-time transmission electron microscopy (TEM) videos. Traditional methods struggled to track the motion of individual defects such as dislocations due to high-dimensional data and noisy imaging artifacts. By using convolutional neural networks for automated defect detection and particle filters for robust object tracking across frames, the new approach successfully measured dislocation velocities in deforming steel with high temporal and spatial precision. The framework enables extraction of detailed motion information that was previously difficult to obtain, opening the door to deeper understanding of defect dynamics and their influence on material properties. Such automated and unbiased processing tools are expected to accelerate discoveries in nanoscale materials research.

Journal: Scientific Reports

DOI:10.1038/s41598-022-13878-8