Deep Learning-Enabled Variational Optimization Method for Image Dehazing in Maritime Intelligent Transportation Systems

Table 2

SSIM comparisons (mean ± std) of various dehazed methods on all test images shown in Figure 5.

PSNR
		0.3	0.5	0.1	0.3	0.5	0.1	0.3	0.5
Haze	12.22 ± 2.08	10.59 ± 2.29	8.77 ± 2.08	14.40 ± 2.08	12.78 ± 2.29	10.96 ± 2.08	17.30 ± 2.09	15.72 ± 2.29	13.87 ± 2.08
DCP	11.92 ± 2.25	9.86 ± 2.30	7.53 ± 1.83	13.39 ± 2.42	10.94 ± 2.11	8.87 ± 1.79	15.23 ± 2.11	12.90 ± 1.80	10.96 ± 1.59
RIVD	14.82 ± 1.84	13.32 ± 2.52	11.09 ± 2.36	17.44 ± 2.25	19.92 ± 2.04	21.94 ± 3.86	16.08 ± 4.02	17.64 ± 3.03	19.10 ± 2.53
MSCNN	13.91 ± 1.97	12.26 ± 2.44	10.17 ± 2.23	18.70 ± 1.55	18.08 ± 2.87	15.68 ± 2.90	21.28 ± 1.76	22.70 ± 2.15	22.20 ± 3.91
Ours	17.71 ± 2.02	16.93 ± 3.33	14.10 ± 3.20	19.63 ± 2.13	22.77 ± 2.79	24.27 ± 5.16	18.86 ± 2.91	20.56 ± 2.30	22.29 ± 2.07