Image Reconstruction Based on Progressive Multistage Distillation Convolution Neural Network

Table 5

Comparison of reconstruction results among algorithms under ×2, ×3, and ×4 reconstruction.

Scale	Method	Set5		Set14		B100		Urban100
		PSNR	SSIM	PSNR	SSIM	PSNR	SSIM	PSNR	SSIM
Comparison of ×2 reconstruction results
×2	SRCNN	36.66	0.954	32.42	0.906	31.36	0.888	29.50	0.895
	VDSR	37.53	0.959	33.03	0.912	31.90	0.896	30.76	0.914
	MemNet	37.78	0.960	33.28	0.914	32.08	0.898	31.31	0.919
	CARN	37.76	0.959	33.52	0.917	32.09	0.898	31.92	0.926
	SRMDNF	37.79	0.960	33.32	0.915	32.05	0.898	31.33	0.920
	IDN	37.83	0.960	33.30	0.915	32.08	0.898	31.27	0.920
	DNCL	37.65	0.960	33.18	0.914	31.97	0.897	30.89	0.916
	FilterNet	37.86	0.961	33.34	0.915	32.09	0.899	31.24	0.920
	MRFN	37.98	0.961	33.41	0.916	32.14	0.899	31.45	0.922
	CFSRCNN	37.79	0.959	33.51	0.916	32.11	0.898	32.07	0.927
	LAPAR-B	37.87	0.960	33.39	0.916	32.10	0.898	31.62	0.924
	MADNet-LF	37.85	0.960	33.39	0.916	32.05	0.898	31.59	0.923
	SMSR	38.00	0.960	33.64	0.918	32.17	0.899	32.19	0.928
	ECBSR	37.90	0.962	33.34	0.918	32.10	0.902	31.71	0.925
	HDRN	37.75	0.959	33.49	0.915	32.03	0.898	31.87	0.925
	SCFFN	38.01	0.960	33.52	0.917	32.12	0.899	31.39	0.926
	PMDN	37.95	0.960	33.58	0.918	32.16	0.899	32.10	0.927
Comparison of ×3 reconstruction results
×3	SRCNN	32.75	0.909	29.28	0.821	28.41	0.786	26.24	0.799
	VDSR	33.66	0.921	29.77	0.831	28.82	0.798	27.14	0.828
	MemNet	34.09	0.925	30.00	0.835	28.96	0.800	27.56	0.838
	CARN	34.29	0.925	30.29	0.841	29.06	0.803	28.06	0.849
	SRMDNF	34.12	0.925	30.04	0.837	28.97	0.803	27.57	0.840
	IDN	34.11	0.925	29.99	0.835	28.95	0.801	27.42	0.836
	DNCL	33.95	0.923	29.93	0.834	28.91	0.799	27.27	0.833
	FilterNet	34.08	0.925	30.03	0.837	28.95	0.803	27.55	0.838
	MRFN	34.21	0.927	30.03	0.836	28.99	0.803	27.53	0.839
	CFSRCNN	34.24	0.926	30.27	0.841	29.03	0.803	28.04	0.850
	LAPAR-B	34.20	0.926	30.17	0.839	29.03	0.803	27.85	0.846
	MADNet-LF	34.14	0.925	30.20	0.839	28.98	0.802	27.78	0.844
	SMSR	34.40	0.927	30.33	0.841	29.10	0.805	28.25	0.854
	HDRN	34.24	0.924	30.23	0.840	28.96	0.804	27.93	0.849
	SCFFN	34.29	0.926	30.27	0.841	29.04	0.803	27.98	0.848
	PMDN	34.36	0.926	30.29	0.840	29.07	0.804	28.15	0.852
Comparison of ×4 reconstruction result
×4	SRCNN	30.48	0.863	27.49	0.750	26.90	0.710	24.52	0.722
	VDSR	31.35	0.884	28.01	0.767	27.29	0.725	25.18	0.752
	MemNet	31.74	0.889	28.26	0.772	27.40	0.728	25.50	0.763
	CARN	32.13	0.894	28.60	0.781	27.58	0.735	26.07	0.784
	SRMDNF	31.96	0.893	28.35	0.777	27.49	0.734	25.68	0.773
	IDN	31.82	0.890	28.25	0.773	27.41	0.730	25.41	0.763
	DNCL	31.66	0.887	28.23	0.772	27.39	0.728	25.36	0.761
	FilterNet	31.74	0.890	28.27	0.773	27.39	0.729	25.53	0.768
	MRFN	31.90	0.892	28.31	0.775	27.43	0.731	25.46	0.765
	CFSRCNN	32.06	0.892	28.57	0.780	27.53	0.733	26.03	0.782
	MADNet-LF	32.01	0.892	28.45	0.778	27.47	0.733	25.77	0.775
	LAPAR-B	31.94	0.892	28.46	0.778	27.52	0.734	25.85	0.777
	SMSR	32.12	0.893	28.55	0.781	27.55	0.735	26.11	0.787
	FDIWN-M	32.17	0.894	28.55	0.781	27.58	0.736	26.02	0.784
	ECBSR	31.92	0.895	28.34	0.782	27.48	0.739	25.81	0.777
	HDRN	32.23	0.896	28.58	0.781	27.53	0.737	26.09	0.787
	SCFFN	32.18	0.895	28.56	0.781	27.54	0.735	26.01	0.783
	PMDN	32.27	0.895	28.58	0.781	27.54	0.735	25.98	0.784