Fig. 2
From: Deep learning revealed statistics of the MgO particles dissolution rate in a CaO–Al2O3–SiO2–MgO slag
The atU-Net particle prediction accuracy for an exemplary video data-set at 1450 °C. (a) Grayscale histogram depicts the intensity area of the MgO-particle (yellow) and the slag (pink) for this selected video. (b) Three representative dissolution times that qualitatively show the accuracy of particle segmentation utilizing the developed atU-Net model. The segmented MgO-particle attained from the atU-net segmentation are highlighted in red. The abbreviation “Dia.” stands for particle diameter. (c) Illustrates the manual evaluation for the same dissolution time. The particles are indicated by the manually drawn particle boundary in green. (d) The evaluated diameter evolution over time is shown for the deep learning atU-Net model (red dots) and the manual evaluation (blue dots), with 1398- and 32-time-steps, respectively.
