Table 3 Multiple methods were used to acquire PSNR and SSIM measurements at varying CT and X-Ray noise levels (blind and specified noise levels); the best results are highlighted in bold

Specific Noise

BM3D [20]

21.99

0.3778

18.63

0.2671

16.08

0.1381

WNNM [26]

21.92

0.3669

18.23

0.2356

16.09

0.1345

NLM [27]

24.61

0.7290

22.32

0.6431

20.29

0.5430

BLS-GSM [28]

30.60

0.8581

29.09

0.8146

27.99

0.7766

RED-CNN [29]

30.46

0.8878

23.98

0.7109

18.19

0.6691

Autoencoder(ANN) [30]

30.30

0.8091

22.11

0.7209

21.34

0.6788

CNN and Wavelets [22]

30.21

0.8912

23.12

0.7217

21.87

0.6823

Non-local means [31]

28.71

0.8371

23.28

0.7154

18.76

0.6544

Blind Noise

Freq. Domain FFT [32]

29.85

0.8822

22.91

0.6992

17.21

0.6141

Adapt. Tensor, PCA [33]

29.46

0.8731

22.74

0.6963

18.31

0.6129

Coeff. Driven Variation [46]

28.75

0.8563

22.41

0.6871

17.78

0.5821

Phase-preserving [34]

28.13

0.8429

22.09

0.6776

17.13

0.6011

Optimal Weight [35]

30.17

0.8879

23.08

0.7050

-

-

Wavelet and Sparse [47]

30.23

0.8078

22.52

0.7053

-

-

DMF-Net(proposed)

31.03

0.8896

29.95

0.8637

29.19

0.8442

DMF-Net(proposed)

29.30

0.8582

28.37

0.8270

27.53

0.8020