EURASIP Journal on Image and Video Processing

Table 2 Comparison of various denoising methods by SSIM (%)

From: Magnitude-phase of the dual-tree quaternionic wavelet transform for multispectral satellite image denoising

Methods	Noise variance
	15	20	25	30	40	50
Image 1
Noisy image	83.96	76.70	71.03	60.24	53.26	44.57
VisuShrink	84.63	65.96	66.45	53.19	55.87	36.55
NeighShrink	73.40	76.50	74.66	71.38	62.47	58.13
BiShrink	83.09	81.57	76.71	88.85	72.04	74.51
Proposed algorithm	86.53	85.57	83.61	92.80	74.93	75.78
Image 2
Noisy image	89.28	82.84	77.25	66.08	58.24	48.22
VisuShrink	83.87	70.32	70.47	64.97	58.80	57.56
NeighShrink	92.52	87.06	84.67	79.73	71.56	82.86
BiShrink	91.53	88.23	88.27	75.59	83.88	82.15
Proposed algorithm	93.81	90.02	88.41	87.83	86.60	86.26
Image 3
Noisy image	70.12	59.94	52.42	39.91	32.68	24.97
VisuShrink	61.30	53.84	68.96	52.23	51.75	48.07
NeighShrink	80.80	75.50	71.96	56.63	54.79	52.36
BiShrink	83.92	80.93	73.37	64.72	64.55	59.46
Proposed algorithm	84.36	84.62	77.41	69.67	74.01	68.02

Wavelet transform: dual-tree quaternionic wavelet transform with five levels.

Back to article page