An Adaptive Boosting Algorithm for Image Denoising

Table 1

Comparison among the denoising results (PSNR, SSIM, and DSI) of BM3D and the proposed algorithm.

BM3D(, )
	Forman			Lena			House
		SOS	OURS		SOS	OURS		SOS	OURS
10	37.24	37.24	37.24	35.87	35.87	35.87	36.65	36.65	36.65
	0.9375	0.9368	0.9368	0.9149	0.9141	0.9154	0.9187	0.9163	0.9220
	2.1859	2.3486	1.7016	3.1644	3.3521	2.8190	2.7045	2.8983	2.0538
20	34.43	34.43	34.50	32.98	32.99	33.01	33.67	33.69	33.67
	0.9061	0.9085	0.9067	0.8764	0.8751	0.8771	0.8684	0.8657	0.8663
	3.3650	3.7503	3.6073	4.7598	5.2280	5.0152	4.4826	4.9408	5.0619
25	33.44	33.46	33.60	32.02	32.04	32.07	32.79	32.80	32.85
	0.8922	0.8978	0.8983	0.8599	0.8591	0.8619	0.8566	0.8555	0.8558
	3.6520	4.2599	4.3283	5.4374	6.1221	6.2237	4.9677	5.5707	5.9420
50	30.04	30.13	30.18	28.78	28.81	29.04	29.38	29.42	29.79
	0.8310	0.8453	0.8454	0.7860	0.7926	0.8014	0.7998	0.8057	0.8159
	4.7899	5.3656	6.6166	10.191	10.757	11.867	7.7156	8.3724	9.3492
	Fingerprint			Pepper			Average
		SOS	OURS		SOS	OURS		SOS	OURS
10	32.46	32.46	32.51	34.58	34.64	34.59	35.36	35.37	35.37
	0.9690	0.9690	0.9682	0.9274	0.9267	0.9264	0.9336	0.9326	0.9338
	0.7034	0.7205	0.4263	2.5305	2.6643	2.0317	2.2577	2.3968	1.8065
20	28.81	28.81	28.82	31.23	31.27	31.27	32.22	32.24	32.25
	0.9305	0.9281	0.9308	0.8849	0.8855	0.8859	0.8933	0.8926	0.8934
	6.2994	6.4851	5.3902	5.7494	6.1046	4.8925	4.9312	5.3018	4.7934
25	27.70	27.70	27.69	30.15	30.16	30.19	31.22	31.23	31.28
	0.9121	0.9072	0.9122	0.8667	0.8678	0.8681	0.8775	0.8775	0.8793
	10.310	10.657	10.260	7.6232	8.1311	7.0077	6.3980	6.9482	6.7523
50	24.34	24.36	24.53	26.32	26.38	26.83	27.77	27.82	28.07
	0.8253	0.8143	0.8367	0.7700	0.7797	0.7952	0.8024	0.8075	0.8189
	35.098	34.955	27.840	18.296	18.521	18.802	15.218	15.594	14.895