Fig. 3: Transfer learning performance. | Nature Communications

Fig. 3: Transfer learning performance.

From: Visualizing nexus of porous architecture and reactive transport in heterogeneous catalysis by deep learning computer vision and transfer learning

Fig. 3

a Learning curves for Generators 1 and 2. b Statistical information, and c Visual information for original and reconstructed 3D models. d Predicted results and error for the dimensionless concentration field (cid) using transfer model. e Predicted results for normalized effective reaction rate (Rnorm) using transfer model. f Heat map for transfer Generator 2. g Experimental setup for electrochemical conversion of K3[Fe(CN)6] on foam nickel. h Nickel foam electrode and representative images, along with the normalized concentration field (cid) at the fluid-solid interface at different Péclet number (Pe) and a fixed Damköhler number (Da) of 1.2. i Comparative analysis between tested and predicted Rnorm under different Pe and fluid directions. Source data are provided as a Source Data file.

Back to article page