Race against the Machine: can deep learning recognize microstructures as well as the trained human eye?. (1st March 2021)

• Record Type: Journal Article
• Authors: Larmuseau, Michiel; Sluydts, Michael; Theuwissen, Koenraad; Duprez, Lode; Dhaene, Tom; Cottenier, Stefaan
• Is Part Of: Scripta materialia. Number 193(2021)
• Page Start: 33
• Page End: 37
• Publication Date: 2021-03-01
• Subjects: Image analysis -- Steels -- Modeling -- Scanning electron microscopy (SEM); Materials -- Periodicals; Metallurgy -- Periodicals; Metalen; Legeringen; Materiaalkunde; Metals, metalworking and machinery industries; Metals; Electronic journals; 620.11
• DOI: 10.1016/j.scriptamat.2020.10.026
• Languages: English
• ISSNs: 1359-6462

Abstract

Graphical abstract: Abstract: The promising results of deep learning in image recognition suggest a huge potential for microscopic analyses in materials science. One major challenge for its adoption in the study of materials is the limited number of images that are available to train models on. Herein, we present a methodology to create accurate image recognition models with small datasets. By explicitly taking into account the magnification and by introducing appropriate transformations, we incorporate as many insights from material science in the model as possible. This allows for a highly data-efficient training of complex deep learning models. Our results indicate that a model trained with the presented methodology is able to outperform human experts.