Shadow Removal on ISTD original (test) benchmark leaderboard · SOTA2 Research

Shadow Removal on ISTD original (test)

4.13RMSE (Shadow)ShadowDiffusion

Evaluation Results

Method
ShadowDiffusion2022.12	4.13	4.14	4.12	40.15	0.994	33.7	0.977	32.33	0.969
DHAN2022.12	7.49	5.3	5.66	35.53	0.988	31.05	0.971	29.11	0.954
BMNet2022.12	7.6	4.59	5.02	35.61	0.988	32.8	0.976	30.28	0.959
Auto-exposure Fusion Network2021.03	7.77	5.56	5.92	—	—	—	—	—	—
Fu et al.2022.12	7.91	5.51	5.88	34.71	0.975	28.61	0.88	27.19	0.945
DHAN2021.03	8.14	6.04	6.37	—	—	—	—	—	—
Zhu et al.2022.12	8.29	4.55	5.09	36.95	0.987	31.54	0.978	29.85	0.96
DSC2022.12	8.72	5.04	5.59	34.64	0.984	31.26	0.969	29	0.944
DSC2021.03	9.76	6.14	6.67	—	—	—	—	—	—
ST-CGAN2022.12	9.99	6.05	6.65	33.74	0.981	29.51	0.958	27.44	0.929
ST-CGAN2021.03	10.33	6.93	7.47	—	—	—	—	—	—
G2R2022.12	10.72	7.55	7.85	31.63	0.975	26.19	0.967	24.72	0.932
DC-ShadowNet2022.12	11.43	5.81	6.57	31.69	0.976	28.99	0.958	26.38	0.922
MaskShadow-GAN2021.03	12.67	6.68	7.41	—	—	—	—	—	—
MaskShadow-GAN2022.12	12.67	6.68	7.41	—	—	—	—	—	—
Gong et al.2021.03	14.98	7.29	8.53	—	—	—	—	—	—
Guo et al.2022.12	18.65	7.76	9.26	27.76	0.964	26.44	0.975	23.08	0.919
Guo et al.2021.03	18.95	7.46	9.3	—	—	—	—	—	—
Input Image2022.12	32.1	7.09	10.88	22.4	0.936	27.32	0.976	20.56	0.893
Input Image2021.03	32.12	7.19	10.97	—	—	—	—	—	—