The physical properties of a material, such as hardness, ductility, toughness, strength, and corrosion resistance, can be affected by its microstructure. These properties help determine how the material will perform in a given application. Experts conduct the microstructure analysis to investigate the failures and supplement the performance of materials.
What is Material Microstructure Analysis?
Microstructure analysis is one of the core techniques in materials science. It ranges from determining certain parameters like grain size, coating thickness, porosity, and pore structure to characterizing multi-component systems or evaluating degradation or failure mechanisms.
In conducting microstructure analysis, experts perform a combination of techniques to provide physical and chemical information about the material with sub-micron resolution. The analysis techniques often complement other methods, such as physical, mineralogical, and chemical approaches. Some areas where microstructural analysis can be utilized to assess and develop products include failure mechanisms, environmental response, and material performance.
In quantitative microstructure analysis, the structural statistics are obtained from microscopic images. The microscopic imaging systems visualize a material's structure information at multiple levels, from the nanoscale to the mesoscale. However, due to the complex and diverse nature of microstructure, humans or artificial intelligence find it hard to perform this procedure.
More Accessible and Faster Microstructure Analysis
At the Korea Institute of Materials Science (KIMS), a research team has developed a technology that can automatically identify and quantify materials microstructure from microscopic images. They are led by Dr. Se-Jong and Dr. Juwon Na of the Materials Data Management Center in the Materials Digital Platform Division in collaboration with a group of experts led by Professor Seungchul Lee of POSTECH.
In this study, the experts effectively integrated human and AI capabilities to develop an integrated framework for quantitative microstructure analysis. This technology allows AI to carry out microstructure segmentation using only a single microscopic image and its corresponding scribble annotation by domain experts. The AI interacts with humans by actively requesting scribble annotation from experts to enhance the performance and reliability of the model.
After extensive experiments, the researchers confirmed the universality of the human-AI collaboration framework and its applicability to a wide range of microstructures, materials, and microscopic imaging systems.
Previous research required collecting large amounts of dense annotation. In this new study, the annotation cost was greatly reduced as the dense annotation was replaced by scribble annotation that can be drawn with a pen or mouse.
This technology developed by the researchers will be incorporated into the Automated Microstructure Quantitative Analysis System (TIMS) currently being developed by KIMS, making it easy to use by general researchers.
According to Dr. Juwon Na, this study is the result of improving the existing subjective and time-consuming quantitative microstructure analysis into an objective and automated technique. Professor Seungchul Lee added that their expert-interacting framework can be a core analysis technology in industry and research. Through this technology, they are confident they can reduce the time and cost of new materials research and development with significantly improved reliability.
RELATED ARTICLE: Novel Approach on Imaging by Combining Two Distinct Microscopy Technology Cells Show How Cells Work
Check out more news and information on Microscopy in Science Times.
