October 4, 2023


It's Your Education

Conquering the limitations of scanning electron microscopy with AI — ScienceDaily

What if a super-resolution imaging method utilized in the most recent 8K premium TVs is applied to scanning electron microscopy, necessary devices for elements exploration?

A joint analysis crew from POSTECH and the Korea Institute of Supplies Science (KIMS) applied deep learning to the scanning electron microscopy (SEM) to produce a super-resolution imaging method that can transform a very low-resolution electron backscattering diffraction (EBSD) microstructure illustrations or photos acquired from conventional examination equipment into tremendous-resolution images. The results from this analyze had been just lately revealed in the npj Computational Supplies.

In modern-day-day elements investigate, SEM pictures participate in a important role in acquiring new elements, from microstructure visualization and characterization, and in numerical materials actions assessment. Nevertheless, acquiring significant-top quality microstructure picture data could be exhaustive or very time-consuming owing to the hardware limitations of the SEM. This might have an effect on the accuracy of subsequent substance examination, and hence, it is paramount to triumph over the technological limitations of the gear.

To this, the joint investigation workforce developed a a lot quicker and much more exact microstructure imaging procedure applying deep finding out. In unique, by applying a convolutional neural network, the resolution of the current microstructure graphic was increased by 4 moments, 8 moments, and 16 situations, which minimizes the imaging time up to 256 instances in contrast to the regular SEM method.

In addition, super-resolution imaging verified that the morphological aspects of the microstructure can be restored with superior accuracy by way of microstructure characterization and finite ingredient analysis.

“Through the EBSD method created in this examine, we anticipate the time it normally takes to develop new resources will be substantially minimized,” stated Professor Hyoung Seop Kim of POSTECH who led the analysis.

This investigate was executed with the assistance from the Mid-job Researcher Software of the National Investigate Basis of Korea, the AI Graduate College Software of the Institute for Details & Communications Technology Promotion (IITP), and Phase 4 of the Mind Korea 21 Application of the Ministry of Education, and with the aid from the Korea Resources Research Institute.

Tale Source:

Elements furnished by Pohang College of Science & Technology (POSTECH). Note: Articles may possibly be edited for fashion and duration.