Skip to main content

Table 4 Quantitative analyses on the restoration results of the 1 mm image data in terms of average PSNR and SSIM scores using different methods

From: An automatic restoration framework based on GPU-accelerated collateral filtering in brain MR images

Type

Noise

LMMSE

IBF

Proposed

PSNR

SSIM

PSNR

SSIM

PSNR

SSIM

Normal

1%

34.97

0.8513

38.04

0.9766

39.17

0.9816

3%

33.34

0.8626

33.28

0.8924

34.79

0.9027

5%

30.65

0.8042

30.51

0.8239

32.42

0.8538

7%

28.65

0.7535

28.60

0.7711

30.53

0.7959

9%

27.03

0.7127

27.04

0.7309

29.13

0.7585

MS

1%

35.09

0.8512

38.16

0.9767

39.29

0.9830

3%

33.48

0.8628

33.40

0.8928

34.87

0.9006

5%

30.79

0.8046

30.64

0.8244

32.51

0.8540

7%

28.79

0.7541

28.73

0.7715

30.62

0.7958

9%

27.17

0.7135

27.16

0.7315

29.21

0.7587