Table 2.

Mean and standard deviation of the edge preservation, texture preservation, mean to standard deviation ratio, contrast to noise ratio, and equivalent number of look for 17 spectral domain optical coherence tomography retinal images by the use of three dimensional CWDL,[47] Tikhonov,[48] MSBTD,[46] K-SVD,[47] K-SVD based DCUT,[15] and proposed method

	Original	3D CWDL[47]	Tikhonov[48]	MSBTD[46]	K-SVD[47]	K-SVD based DCUT[15]	Proposed method
Mean±STD (EP)[49]	1±0	0.91±0.22	0.71±0.45	0.68±0.59	0.63±0.33	0.96±0.02	0.95±0.01
Mean±STD (TP)[49]	1±0	0.41±0.52	0.18±0.85	0.12±0.87	0.32±0.25	0.75±0.05	0.96±0.02
Mean±STD (MSR)[50]	3.20±0.46	14.45±4.85	7.64±0.63	14.76±4.75	11.22±2.77	14.36±3.85	29.33±20.18
Mean±STD (CNR)[51]	1.27±0.43	7.31±3.63	3.26±0.22	4.76±1.54	4.11±1.23	7.81±3.54	31.07±19.78
Mean±STD (ENL)[49]	38.09±4.35	2323.65±43.54	74.34±4.32	783.68+67.65	132.01±7.32	1983.08±39.42	69.51±138.12
Mean±STD (PSNR)[52]	-	-	23.67±0.96	26.46±1.72	26.13±1.70	17.96±1.08	15.55±1.75

The best result in the table is shown in bold. STD – Standard deviation; EP – Edge preservation; TP – Texture preservation; MSR – Mean to standard deviation ratio; CNR – Contrast to noise ratio; ENL – Equivalent number of look; 3D – Three dimensional; CWDL – Complex Wavelet-based Dictionary Learning; MSBTD – Multi-scale Sparsity-based Tomographic Denoising; K-SVD – K-Singular Value Decomposition; DCUT – Digital Curvelet Transform