Table 2 The FID, LPIPS, SSIM, and PSNR of Imagest and several baseline methods

From: Imagest: a style conditioned art painting synthesis and generation networks

Model

Imagest

VQGAN

Stable Diffusion 1–4

Stable Diffusion 1–5

DALL.E 2

FID↓

10.62

12.89

12.30

10.87

10.03

LPIPS↓

0.244

0.280

0.252

0.249

0.225

SSIM ↑

0.882

0.835

0.864

0.879

0.892

PSNR ↑

27.5

24.3

25.8

26.9

27.1

  1. Bold values indicate the best performance under each metric, while underlined values denote the second-best.