The Scientific World Journal / 2014 / Article / Tab 4

Research Article

A “Salt and Pepper” Noise Reduction Scheme for Digital Images Based on Support Vector Machines Classification and Regression

Table 4

Table results in mean MSSIM for different impulse noise reduction methods. The mean is obtained for different noise ratios in 10 test images. Total column reflects the total mean for all ratios and images.

MethodNoise percentage
10 20 30 40 50 60 70 80 90Total

ACWM [19]0.9300.8760.7340.5020.2770.1410.0690.0330.0140.397
AMF [12]0.9220.9090.8840.8470.8000.7420.6530.4660.1390.707
DBAIN [20]0.9780.9510.9180.8770.8250.7610.6770.5700.4270.776
DSFIRE [21]0.9610.9050.7830.5750.3480.1830.0880.0400.0160.433
FIDRM [22]0.9770.9480.9120.8700.8210.7670.6960.5970.4560.783
MMEM [14] 0.9430.9180.8930.8680.8370.8010.754 0.693 0.596 0.811
NAFSM [23]0.9680.9380.9040.8680.8280.7840.7320.6650.5380.803
PSMF [17]0.8560.8360.8020.7540.6710.5090.2720.0450.0190.529
SDROM [5]0.9250.8830.8290.7540.6450.5080.3430.1790.0670.570
MDBUTMF [24] 0.979 0.956 0.930 0.897 0.8510.7650.6080.3840.1780.728
SVM-M1 (32-40) 0.9780.9500.9120.8590.8020.7230.6170.3660.0770.698
SVM-M1 (64-40) 0.9780.9500.9120.8600.8030.7250.6190.3690.0770.699
SVM-M2 (32-40) 0.9780.9540.9260.8930.8530.807 0.7520.6850.5850.826
SVM-M2 (64-40) 0.9780.9540.9250.8920.8520.8060.7510.6840.5840.825
SVM-R (32-40) 0.9760.9530.9260.8940.854 0.8030.7350.6360.4780.806
SVM-R (64-40) 0.9760.9530.9260.8940.854 0.8030.7350.6360.4790.806