Clearer and Better Focused SEM Images
Clearer and Better Focused SEM Images
  • Reporter Lee Seung-ah
  • 승인 2021.06.27 19:41
  • 댓글 0
이 기사를 공유합니다

▲Improving quality of SEM images using AI
▲Improving quality of SEM images using AI


With the onset of the Fourth Industrial Revolution, AI has recently been utilized in smartphone cameras, providing functions such as auto-focusing, face recognition, and 100x zoom. It has also been applied to the research and development of new materials.
A joint research team from POSTECH and the Korea Institute of Materials Science (KIMS) has applied deep learning to the scanning electron microscopy (SEM) system, which is an essential material analysis equipment, to detect and improve the quality of SEM images without human intervention. The findings were recently published in Acta Materialia, the most authoritative academic journal in the field of metal materials.
SEM is one of the most advanced material analysis equipment used when determining the correlation between the microstructural and physical, chemical, and mechanical properties of materials by acquiring their microstructural images. However, the operator’s high proficiency and detailed manipulation are essential for obtaining high-quality, clear SEM images–otherwise, it can lead to low-quality microscopic images. The quality of these images must be improved since they directly affect the subsequent material analysis process.
As such, the joint research team developed a technique based on deep learning to automatically detect and improve the quality of these images. The developed technology is based on a multi-scale deep neural network and it demonstrated that it can improve the quality of microstructural images on blind settings without any prior knowledge or assumptions of the degree of blurring on the level of image degradation. In addition, the research team proposed a technique that allows AI to learn not only how but also where to refocus in non-uniformly defocused images.
“We expect to reduce the development cost and time of new materials by automating the imaging process of SEM, which is widely used for research and development of new materials,” remarked Professor Seungchul Lee (ME), who conducted the research.
This research was conducted with support from the Mid-career Researcher Program, the Priority Research Centers Program, the AI Graduate School Program of the Institute for Information & Communication Technology Promotion, and the Korea Institute of Materials Science.