Cardiac Magnetic Resonance Images Superresolution via Multichannel Residual Attention Networks

Table 2

Average PSNR/SSIM of various SSIR methods; the best and second-best results are in bold and italics.

Algorithm	Scale	Set 5		Set 14		Urban 100
		PSNR	SSIM	PSNR	SSIM	PSNR	SSIM
Bicubic [6]	2×	33.69	0.931	30.25	0.870	26.88	0.841
A+ [27]	2×	36.60	0.955	32.32	0.906	29.25	0.895
SCN [28]	2×	36.58	0.954	32.35	0.905	29.52	0.897
SRCNN [9]	2×	36.72	0.955	32.51	0.908	29.53	0.896
FSRCNN [10]	2×	37.05	0.956	32.66	0.909	29.88	0.902
VDSR [11]	2×	37.53	0.959	33.05	0.913	30.77	0.914
DRCN [12]	2×	37.63	0.959	33.06	0.912	30.76	0.914
LapSRN [14]	2×	37.52	0.959	33.08	0.913	30.41	0.910
DRRN [29]	2×	37.74	0.959	33.23	0.914	31.23	0.919
MCRN (ours)	2×	37.88	0.961	33.63	0.919	31.27	0.919
Bicubic [6]	3×	30.41	0.869	27.79	0.775	24.46	0.735
A+ [27]	3×	32.62	0.909	29.15	0.820	26.05	0.799
SCN [28]	3×	32.62	0.908	29.16	0.818	26.21	0.801
SRCNN [9]	3×	32.78	0.909	29.32	0.823	26.25	0.801
FSRCNN [10]	3×	33.18	0.914	29.37	0.824	26.43	0.808
VDSR [11]	3×	33.67	0.921	29.78	0.832	27.14	0.829
DRCN [12]	3×	33.83	0.922	29.77	0.832	27.15	0.828
LapSRN [14]	3×	33.82	0.922	29.87	0.832	27.07	0.828
DRRN [29]	3×	34.03	0.924	29.96	0.835	27.53	0.764
MCRN (ours)	3×	33.99	0.925	30.28	0.844	27.27	0.830
Bicubic [6]	4×	28.43	0.811	26.22	0.715	23.14	0.658
A+ [27]	4×	30.32	0.860	27.34	0.751	24.34	0.721
SCN [28]	4×	30.41	0.863	27.39	0.751	24.52	0.726
SRCNN [9]	4×	30.50	0.863	27.52	0.753	24.53	0.725
FSRCNN [10]	4×	30.72	0.866	27.61	0.755	24.62	0.728
VDSR [11]	4×	31.35	0.883	28.02	0.768	25.18	0.754
DRCN [12]	4×	31.54	0.884	28.03	0.768	25.14	0.752
LapSRN [14]	4×	31.54	0.885	28.19	0.772	25.21	0.756
DRRN [29]	4×	31.68	0.888	28.21	0.772	25.44	0.764
MCRN(ours)	4×	31.67	0.887	28.45	0.783	25.25	0.759