Table 6 Comparison of evaluation metrics on the landscape painting dataset (ink wash style)

From: Color reconstruction of ancient paintings based on dual pooling attention and pixel adaptive convolution

Model

SSIM ↑ 

Gram Loss ↓ 

CAP-VSTNet

0.793

0.00363

CCPL

0.686

0.01038

ArtFlow

0.664

0.00212

PCA-KD

0.729

0.02956

PhotoWCT2

0.715

0.06572

DTP

0.742

0.00543

Ours

0.806

0.00203

  1. The bold values indicate the best performance in each table.
Back to article page