Scientific Reports

Table 2 Quantitative comparison on LOL (V1 & V2) dataset in terms of PSNR, SSIM, MSE, AB, LPIPS, and NIQE.

From: Low-light image enhancement via adaptive frequency decomposition network

Method	LOL-V1						LOL-V2
Method	PSNR\(\uparrow \)	SSIM\(\uparrow \)	MSE\(\downarrow \)	AB \(\uparrow \)	LPIPS\(\downarrow \)	NIQE\(\downarrow \)	PSNR\(\uparrow \)	SSIM\(\uparrow \)	MSE\(\downarrow \)	AB \(\uparrow \)	LPIPS\(\downarrow \)	NIQE\(\downarrow \)
LLNet¹⁴	28.1241	0.5129	101.5715	64.3940	0.3491	8.0910	27.9700	0.5269	104.5375	61.8057	0.3477	8.2731
MBLLEN¹⁷	28.0733	0.7863	103.3022	100.9130	0.2111	4.3592	27.9410	0.6423	104.7113	80.2747	0.2737	4.8441
Retinex-Net¹⁶	28.0565	0.4191	101.8668	107.7655	0.4443	8.8792	27.9991	0.4006	103.2850	101.5877	0.5429	9.4274
Zero-DCE¹²	27.8289	0.6626	108.6866	72.0175	0.3142	7.7667	28.1533	0.5735	101.2444	65.0908	0.3111	8.0578
Zero-DCE++²⁰	28.0184	0.5619	103.8069	73.5041	0.3208	7.8963	28.1504	0.5718	101.3338	64.6287	0.3126	8.0448
EnlightenGAN¹³	27.8051	0.7319	108.6318	94.7671	0.2968	6.4889	28.1550	0.6754	99.7077	82.7959	0.3088	5.0888
Kind¹⁸	27.9912	0.7748	104.1301	103.2690	0.1641	4.7101	28.0865	0.7145	101.8748	88.9531	0.3770	5.5608
Kind++¹⁹	28.1693	0.7696	102.6426	109.1153	0.1853	4.7663	28.1095	0.7695	100.9257	107.8107	0.2165	4.6714
R2RNet⁶⁰	28.1359	0.74317	100.6120	124.3166	0.2851	4.0877	28.0023	0.7747	84.6253	121.3354	0.2724	4.9576
SCI²⁹	27.9048	0.5219	105.9507	69.2669	0.3393	7.8767	27.9600	0.5335	104.9082	63.76475	0.3079	8.0461
RUAS²⁸	28.0313	0.4996	102.7765	94.13435	0.2701	6.3401	27.9219	0.4875	105.2823	85.2675	0.3097	6.5331
HFMNet²¹	28.5593	0.8147	93.2648	112.6408	0.1268	4.5616	28.8586	0.8273	88.1210	104.6577	0.1406	4.4929
Ours	28.9181	0.8238	88.9780	117.1915	0.1436	4.0775	29.3200	0.8373	79.1373	100.2312	0.1397	4.3192

The up or down arrows represent the metrics go direct in good. best and second best results are marked in bold and italic respectively.

Back to article page

Search

Advanced search

Quick links