Abstract
Automated, reliable, and objective microstructure inference from micrographs is essential for a comprehensive understanding of process-microstructure-property relations and tailored materials development. However, such inference, with the increasing complexity of microstructures, requires advanced segmentation methodologies. While deep learning offers new opportunities, an intuition about the required data quality/quantity and a methodological guideline for microstructure quantification is still missing. This, along with deep learning’s seemingly intransparent decision-making process, hampers its breakthrough in this field. We apply a multidisciplinary deep learning approach, devoting equal attention to specimen preparation and imaging, and train distinct U-Net architectures with 30–50 micrographs of different imaging modalities and electron backscatter diffraction-informed annotations. On the challenging task of lath-bainite segmentation in complex-phase steel, we achieve accuracies of 90% rivaling expert segmentations. Further, we discuss the impact of image context, pre-training with domain-extrinsic data, and data augmentation. Network visualization techniques demonstrate plausible model decisions based on grain boundary morphology.
Highlights
Automated, reliable, and objective microstructure inference from micrographs is essential for a comprehensive understanding of process-microstructure-property relations and tailored materials development
The task is framed as binary segmentation which we address with supervised learning of convolutional neural networks (CNN)
Unifying different methodologies such as specimen preparation, image acquisition, multimodal data registration, data fusion, deep learning modeling, and network visualization, all described in the “Methods” section, facilitates a holistic approach for microstructure inference
Summary
Reliable, and objective microstructure inference from micrographs is essential for a comprehensive understanding of process-microstructure-property relations and tailored materials development Such inference, with the increasing complexity of microstructures, requires advanced segmentation methodologies. 2D micrographs of different imaging modalities, such as light optical microscopy (LOM) or scanning electron microscopy (SEM), are utilized for microstructure inference Such micrographs’ automated, reliable, and objective segmentation is not established for all desirable material classes. Microstructure recognition tasks, compared to natural images such as ImageNet[8], can be very complex regarding the degree of detail and information density in the images This further impedes the determination of accurate annotations[9] needed for supervised-learning, which may discourage the use of DL, resulting in a lack of representative annotated open-access data sets.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
