Machine learning-based image processing in materials science and engineering: A review

Machine learning (ML) is playing a great role in every sphere of life including the area of materials science and engineering. At times, the experimental and computational data together create a complex scenario for interpretation giving rise to the requirement of ML. The incorporation of ML in the development of material science has a promised future to accelerate the research through automatic data interpretation with the help of different models like classification, Regression, Clustering, etc. Therefore, it becomes quick and easy to analyze the data sets with greater accuracy as compared to the analysis done by human researchers. This paper gives an overview of the applications and applicability of Machine Learning being used in various areas of materials development. The various image processing techniques which can be applied in material science are discussed in detail. In addition to that, the paper gives detail as to how a model can be trained with a smaller number of datasets. Subsequently, the different data ecosystems which can be useful for data collection and preparation before selecting any Machine Learning model are highlighted. Furthermore, the various opportunities and challenges that material scientists have been facing in the development of new materials are also outlined for a better understanding of ML-based image processing in material science and engineering.Copyright (c) 2022 Elsevier Ltd. All rights reserved.Selection and peer-review under responsibility of the 9th International Conference on Advancements and Futuristic Trends in Mechanical and Materials Engineering(AFTMME 2021).