Table 3 Evaluated R2 to assess the synthetic image quality

From: Analyzing microstructure relationships in porous copper using a multi-method machine learning-based approach

Sample

R2

Relative density D

Specific perimeter SP2D

Shape index

Avg. of physical descriptors

cGAN

DDPM

cGAN

DDPM

cGAN

DDPM

cGAN

DDPM

HPA

0.33

0.76

0.75

0.86

0.72

0.89

0.60

0.84

HPB

0.28

0.72

0.69

0.83

0.34

0.92

0.44

0.82

NPC

0.88

0.86

0.87

0.97

0.34

0.92

0.87

0.90

  1. Comparison of the different R2 values for the relative density D, the specific perimeter SP2D,  the shape index and their average, in relationship to the different samples (HPA, HPB, NPC) and microstructure prediction models (cGAN, DDPM). The analysis indicates that the DDPM provides better results than the cGAN model.
Back to article page