Deep Learning-based Computer Vision for Radiation Defect Analysis: from Static Defect Segmentation to Dynamic Defect Tracking

Over the past several decades, immense research has been dedicated to the understanding of radiation effects on materials, particularly driven by the needs of the nuclear power industry. Yet, researchers have been unable to reach a unified and comprehensive predictive insight to the response of reactor structural materials under energetic particle irradiation. This is in part due to the lack of satisfactory TEM image processing tools capable of performing reliable, fast, and reproducible radiation defect detection and tracking, leaving experimental TEM data, especially new in situ irradiation TEM video data, underexplored.