Fig. 4: Accuracy comparison for different methods. | Nature Communications

Fig. 4: Accuracy comparison for different methods.

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

Fig. 4

a–d Conditions with similar quantitative structural features (QSFs) but different anisotropy factor gi. a Normalized concentration (cid) distribution and b Normalized effective reaction rate (Rnorm) variations for axis anisotropic architecture; c cid distribution and d Rnorm variations for diagonal anisotropic architecture. e, f Conditions with the same QSFs and gi but different flow directions. e cid distribution and f Rnorm variations under x+ fluid for structure with diagonal textures on x-z plane. g cid distribution and h Rnorm variations under y+ fluid for structure with diagonal textures on x-z plane. HA and HB were hypothetical structures with similar QSFs but different gi values. HC was the hypothetical anisotropic structure with diagonal textures. Source data are provided as a Source Data file.

Back to article page