Fig. 1. Comparison of average values of PNSR and FSIM indexes for Berkeley test dataset images by various algorithms. (a) Comparison result of PNSR values; (b) comparison result of FSIM values

Fig. 2. Comparison of denoising effects of noised MRI brain slices by various algorithms. (a) MRI slice with noise; (b) denoising effect by RM algorithm; (c) denoising effect by BM3D algorithm; (d) denoising effect by WSNM algorithm; (e) denoising effect by WNNM algorithm; (f) denoising effect by proposed algorithm

Fig. 3. Denoising residual components of Lena image (σ= 40). (a) Residual component of RM algorithm; (b) residual component of BM3D algorithm; (c) residual component of proposed algorithm; (d) residual component of WSNM algorithm; (e) residual component of WNNM algorithm

Fig. 4. Comparison of denoising details of Male image in salt and pepper noise with the noise density of 50. Figs. 4(c)-(g) are detail parts of Fig. 4(a). (a) Original Male image; (b) salt and pepper noise image with noise density of 50; (c) result of proposed algorithm; (d) result of WNNM algorithm; (e) result of WSNM algorithm; (f) result of BM3D algorithm; (g) result of RM algorithm

Table 1. Comparison between proposed algorithm and other algorithms on PSNR and FSIM by selecting four images of Boat, Male, Peppers, and Pentagon under different salt and pepper noise densities