Dictionary Learning Based on Nonnegative Matrix Factorization Using Parallel Coordinate Descent

Table 1

PSNR (dB) and SSIM results for different algorithms. In each cell, four groups of denoising results are shown. Top row, Wavelets; second row, NL-means; third row, NMFℓ⁰-H; bottom row, PCDDL.

Input PSNR	Lena		Barbara		Boat		House		Pepper		Average
	PSNR	SSIM	PSNR	SSIM	PSNR	SSIM	PSNR	SSIM	PSNR	SSIM	PSNR	SSIM
	32.91	0.8775	29.61	0.8633	30.28	0.8093	33.05	0.8669	31.20	0.8881	31.41	0.8610
29.97	37.76	0.9370	36.62	0.9583	35.55	0.9205	38.17	0.9357	36.02	0.9485	36.82	0.9400
	29.23	0.9224	33.15	0.9538	35.95	0.9451	33.68	0.9519	35.76	0.9568	33.55	0.9460
	39.39	0.9491	38.69	0.9641	38.08	0.9445	40.10	0.9580	38.63	0.9577	38.98	0.9547
20.12	28.52	0.7982	25.03	0.7168	26.63	0.6988	28.03	0.7856	26.24	0.7902	26.89	0.7579
	33.45	0.8756	31.74	0.8914	31.04	0.8148	34.06	0.8729	32.06	0.8929	32.47	0.8695
	29.61	0.8680	30.82	0.9056	29.17	0.8232	33.93	0.8846	28.22	0.8788	30.35	0.8720
	34.39	0.8791	32.58	0.8870	32.24	0.8376	34.32	0.8714	32.70	0.8839	33.25	0.8718
14.09	25.74	0.7312	22.71	0.6044	24.23	0.6101	24.90	0.7220	23.15	0.7021	24.15	0.6740
	30.14	0.8039	27.52	0.7867	27.69	0.7145	30.44	0.8087	28.52	0.8235	28.86	0.7875
	27.68	0.7870	24.63	0.7411	24.10	0.6578	27.64	0.8170	22.87	0.7604	25.38	0.7527
	31.24	0.7953	28.71	0.7700	28.80	0.7137	31.21	0.7981	29.27	0.7933	29.85	0.7741
8.82	23.47	0.6712	21.07	0.5249	22.32	0.5384	22.51	0.6666	20.56	0.6164	21.99	0.6035
	27.18	0.6876	24.32	0.6402	24.99	0.5938	26.60	0.6767	24.97	0.7114	25.61	0.6619
	21.58	0.6677	19.60	0.5175	20.38	0.5190	21.41	0.6746	19.31	0.6232	20.46	0.6004
	28.38	0.6727	25.26	0.5986	26.08	0.5693	28.37	0.6917	26.27	0.6643	26.87	0.6393