Image-driven discriminative and generative machine learning algorithms for establishing microstructure–processing relationships

• Published in: Journal of applied physics 2020-10, Vol.128 (13)
• Main Authors: Ma, W., Kautz, E. J., Baskaran, A., Chowdhury, A., Joshi, V., Yener, B., Lewis, D. J.
• Format: Article
• Language: English
• Subjects: Algorithms; Applied physics; Best practice; Binary alloys; Datasets; Machine learning; Metallurgical analysis; Microstructure; Nuclear fuels; Object recognition; Performance evaluation; Photomicrographs; Representations; Thermomechanical treatment; Uranium
• ISSN: 0021-8979; 1089-7550
• DOI: 10.1063/5.0013720

We investigate the methods of microstructure representation for the purpose of predicting processing condition from microstructure image data. A binary alloy (uranium–molybdenum) that is currently under development as a nuclear fuel was studied for the purpose of developing an improved machine learning approach to image recognition, characterization, and building predictive capabilities linking microstructure to processing conditions. Here, we test different microstructure representations and evaluate model performance based on the F1 score. A F1 score of 95.1% was achieved for distinguishing between micrographs corresponding to ten different thermo-mechanical material processing conditions. We find that our newly developed microstructure representation describes image data well, and the traditional approach of utilizing area fractions of different phases is insufficient for distinguishing between multiple classes using a relatively small, imbalanced original dataset of 272 images. To explore the applicability of generative methods for supplementing such limited datasets, generative adversarial networks were trained to generate artificial microstructure images. Two different generative networks were trained and tested to assess performance. Challenges and best practices associated with applying machine learning to limited microstructure image datasets are also discussed. Our work has implications for quantitative microstructure analysis and development of microstructure–processing relationships in limited datasets typical of metallurgical process design studies.