Table 2 The performance of different denoising algorithms is evaluated based on Entropy, PSNR, SSIM, FSIM, BRISQUE, PIQE and NIQE metrics for MRI2 image.

From: Detail-preserving denoising of CT and MRI images via adaptive clustering and non-local means algorithm

Method	σ	Entropy	PSNR	SSIM	FSIM	BRISQUE	NIQE	PIQE
Ours	10	6.34	35.58	0.95	0.96	48.07	5.25	55.52
	20	6.36	32.35	0.90	0.93	48.93	4.73	54.53
	30	6.44	30.28	0.86	0.90	44.63	4.72	51.40
	40	6.60	28.64	0.80	0.88	36.57	4.35	53.91
OD-CNN	10	6.12	34.72	0.88	0.93	49.74	5.67	57.01
	20	6.01	31.10	0.86	0.92	49.53	5.59	59.32
	30	6.14	29.38	0.81	0.89	46.98	5.74	56.77
	40	6.13	27.18	0.78	0.87	47.52	5.97	58.78
NDiff	10	5.34	32.06	0.84	0.92	53.50	5.65	60.68
	20	5.43	26.08	0.80	0.87	55.02	5.30	63.42
	30	5.44	26.98	0.79	0.86	53.45	5.83	58.80
	40	5.19	28.22	0.80	0.88	53.46	5.71	60.93
SDPM	10	5.99	28.28	0.81	0.89	53.01	5.78	64.14
	20	6.22	26.45	0.77	0.87	48.43	5.88	64.74
	30	5.24	23.85	0.74	0.86	54.17	5.99	68.09
	40	5.82	23.65	0.81	0.85	45.46	6.14	65.00
NGN	10	6.15	26.98	0.76	0.88	54.74	6.21	59.31
	20	5.29	23.25	0.72	0.86	52.91	6.00	57.90
	30	6.06	23.06	0.75	0.86	62.16	5.96	60.30
	40	5.44	20.16	0.72	0.84	53.51	5.54	64.00
UDLF	10	5.81	30.11	0.85	0.86	50.93	5.79	65.64
	20	6.00	23.91	0.81	0.82	62.60	5.74	60.67
	30	6.06	21.83	0.76	0.81	63.19	6.10	58.74
	40	6.17	21.60	0.73	0.80	53.43	5.80	59.80
CNCL	10	5.58	29.98	0.84	0.78	49.36	5.84	62.39
	20	5.71	25.60	0.81	0. 77	50.44	6.28	65.95
	30	5.75	23.82	0.76	0.77	59.54	5.69	62.60
	40	5.62	18.56	0.73	0.76	54.46	6.50	64.01
BP-MEC	10	5.33	30.64	0.83	0.84	59.14	5.77	63.71
	20	5.49	24.94	0.79	0.82	63.42	5.99	62.22
	30	5.51	23.20	0.75	0.79	63.46	6.19	65.91
	40	5.40	18.23	0.70	0.78	53.49	6.04	66.40
FDCT	10	5.46	29.27	0.79	0.77	60.18	6.15	66.57
	20	5.63	25.60	0.75	0.78	59.60	6.79	67.40
	30	5.69	23.48	0.80	0.73	60.89	6.99	69.31
	40	4.88	19.09	0.76	0.81	63.49	6.74	71.55
NSCT-BT	10	5.42	27.89	0.78	0.75	61.99	6.87	67.19
	20	5.27	25.19	0.73	0.73	62.88	6.98	68.99
	30	5.10	24.81	0.71	0.71	63.89	7.10	70.01
	40	5.03	23.78	0.68	0.69	64.64	7.23	72.22
IDM-PM	10	5.27	26.99	0.76	0.77	62.87	6.93	68.83
	20	5.13	25.81	0.73	0.75	63.99	7.04	69.87
	30	4.98	24.67	0.71	0.73	65.78	7.15	71.88
	40	4.78	22.18	0.69	0.71	66.16	7.20	72.99
DCT-ACO	10	5.06	27.18	0.75	0.79	64.95	6.70	67.74
	20	5.20	23.84	0.74	0.76	59.45	6.99	68.57
	30	5.23	21.99	0.76	0.73	57.12	6.81	74.56
	40	4.62	16.78	0.69	0.70	60.42	7.64	69.08
ANLM	10	4.98	26.93	0.74	0.78	65.89	6.74	67.99
	20	4.79	25.17	0.73	0.75	66.67	6.87	69.89
	30	4.66	23.89	0.71	0.74	69.18	7.01	71.72
	40	4.56	22.19	0.67	0.73	70.18	7.17	72.89
Z-NLM	10	4.87	25.66	0.73	0.77	65.67	6.99	69.99
	20	4.67	23.89	0.69	0.76	66.18	7.23	70.99
	30	4.56	22.29	0.67	0.74	67.78	7.27	71.87
	40	4.23	21.87	0.65	0.71	69.89	7.32	72.19
ONLM	10	4.67	24.89	0.69	0.74	66.98	7.17	70.67
	20	4.45	23.28	0.67	0.72	70.18	7.20	72.89
	30	4.26	22.95	0.64	0.69	71.35	7.25	73.29
	40	4.10	21.89	0.62	0.67	72.97	7.29	74.88

The best performing results are highlighted in bold for clarity.

Back to article page

Table 2 The performance of different denoising algorithms is evaluated based on Entropy, PSNR, SSIM, FSIM, BRISQUE, PIQE and NIQE metrics for MRI2 image.

Search

Quick links