Figure 5
From: Machine learning-based microstructure prediction during laser sintering of alumina
Using computer-generated micrographs to evaluate the accuracy of the RCWGAN-GP in regenerating and predicting microstructure of secondary phase grain growth. The processing parameter are a time series, t/τ = 1.6, 2.4, 3.2, 4.0, 4.8, 5.6, 6.4, 7.2 and 8.0. After training, RCWGAN-GP can regenerate micrographs that have similar microstructure features at the trained time series. We use RCWGAN-GP to predict the microstructure at a new set of time, (t/τ) = 2.0, 2.8, 3.6, 4.4, 5.2, 6.0, 6.8, and 7.6. The predicted micrographs showed the expected microstructure features during the secondary phase grain growth.
