Single Image Dehazing and Edge Preservation Based on the Dark Channel Probability-Weighted Moments

Table 2

Performance comparison of DCPWM in terms of the e, , and with state-of-the-art image dehazing methods for 5 images (values in bold format indicate the best performances).

Techniques	Gate			Tower			Yos1			Yos2			Bridge
	e			e			e			e			e
Yuan and Huang [26]	8.22	0.03	1.88	7.15	0.04	2.88	10.53	0.10	2.05	8.01	0.02	2.08	9.63	0.05	3.15
Ren et al. [31]	4.52	0.56	1.62	5.31	0.75	1.82	7.21	0.32	1.59	5.83	0.11	1.99	4.18	0.10	2.31
Cai et al. [16]	3.78	0.59	1.63	5.17	0.83	1.79	6.98	0.29	1.68	5.29	0.40	2.07	4.57	0.37	2.26
Berman et al. [32]	2.37	1.33	1.77	4.94	0.72	2.03	7.16	0.31	1.89	4.73	0.28	2.21	3.11	0.54	2.76
Choi et al. [33]	1.93	3.12	1.98	5.32	1.52	1.76	6.43	0.88	2.15	3.25	1.10	2.23	2.57	1.00	1.89
Meng et al. [14]	6.52	0.17	1.45	4.13	0.26	1.26	7.80	0.12	1.47	6.51	0.02	1.83	4.32	0.23	2.34
He et al. [7]	5.81	0.20	1.42	4.84	0.17	1.37	7.66	0.13	1.33	6.72	0.02	1.65	3.22	0.25	2.16
DCPWM (proposed)	8.24	0.04	2.2	7.64	0.02	2.92	12.01	0.08	2.49	9.68	0.03	2.52	9.59	0.01	3.26