Table 2 Quantitative results of three supervised metrics (PSNR, SSIM, and LPIPS) and four non-reference metrics (NIQE, NIMA, MUSIQ, and ILNIQE) on the LSRW dataset.

From: CUI-Net: a correcting uneven illumination net for low-light image enhancement

Method

SSIM\(\uparrow\)

PSNR\(\uparrow\)

LPIPS\(\downarrow\)

NIQE\(\downarrow\)

NIMA\(\uparrow\)

MUSIQ\(\uparrow\)

ILNIQE\(\downarrow\)

EnlightGAN

0.5032

17.08

0.3274

3.4691

4.2444

56.3958

27.6677

KinD

0.5039

16.4069

0.3371

3.5432

4.5964

61.2283

27.231

ZeroDCE

0.4773

15.8545

0.3172

3.7209

3.9323

55.8506

30.806

ZeroDCE++

0.4229

12.9252

0.4324

3.7084

3.887

50.335

27.6368

RUAS

0.4324

14.0305

0.3846

4.2598

4.2804

53.9967

36.2296

SCI

0.4474

15.2405

0.322

3.8659

4.276

56.2192

32.1176

Uretinex

0.5453

18.2709

0.2994

4.1814

4.5923

65.8647

29.642

CUI

0.4389

16.564

0.333

3.6239

4.6053

56.6684

29.6124

  1. The best result is highlighted in italic, the second-best is highlighted in bold, and the third-best is highlighted in bolditalic.