242043.fig.004a
(a)
242043.fig.004b
(b)
242043.fig.004c
(c)
242043.fig.004d
(d)
242043.fig.004e
(e)
242043.fig.004f
(f)
242043.fig.004g
(g)
242043.fig.004h
(h)
242043.fig.004i
(i)
242043.fig.004j
(j)
Figure 4: Detail of the denoising medical image with different methods. (a) A slice of an MRI brain image; (b) image with Gaussian white noise of mean 0 and variance 0.01, P S N R = 2 1 . 0 9 8 0 ; (c) denoised with median filter (template: 3 × 3), P S N R = 2 7 . 3 7 7 4 ; (d) denoised with averaging filter (template: 3 × 3), P S N R = 2 7 . 7 7 0 1 ; (e) denoised with Wiener filter (template: 5 × 5), P S N R = 2 7 . 8 0 6 1 ; (f) denoised with TVM ( 𝜆 = 0 . 0 1 , 10 iterations), P S N R = 2 4 . 9 2 0 0 ; (g) results with the original fourth-order PDE model, P S N R = 2 7 . 6 7 3 0 ; (h) denoised (g) with Median filter, P S N R = 2 7 . 7 7 4 3 ; (i) results with the new fourth-order PDE model, P S N R = 2 6 . 9 9 0 2 ; (j) denoised (i) with median filter, P S N R = 2 8 . 4 7 4 6 .