Mathematical Problems in Engineering / 2018 / Article / Tab 2

Research Article

Image Denoising Algorithm Combined with SGK Dictionary Learning and Principal Component Analysis Noise Estimation

Table 2

Denoising indicators on SGK combined with five noise estimation algorithms.

AlgorithmsTime/sPSNR/dBMSE

σ = 2PCA + SGK2.733+0.7330.7331.79943.3852.983
Kurtosis + SGK2.841+0.8410.8414.69943.3782.987
Mode + SGK2.768+0.7680.7680.07043.2223.010
Med + SGK4.063+2.0632.0630.05141.8114.286
Min + SGK0.496−1.5041.5040.03742.1193.993

= 5PCA + SGK5.510+0.5100.5101.82938.3839.436
Kurtosis + SGK5.166+0.1660.1665.86538.3029.614
Mode + SGK4.480−0.5200.5200.09138.1399.981
Med + SGK6.194+1.1941.1940.04738.2519.727
Min + SGK0.674−4.3264.3260.03534.17524.865

= 10PCA + SGK10.342+0.3420.3421.60435.17719.201
Kurtosis + SGK9.576+0.4240.4244.84834.79521.559
Mode + SGK8.809+1.1911.1910.09034.07425.447
Med + SGK10.463+0.4630.4630.05135.14919.869
Min + SGK1.640−18.36018.3600.03728.24497.417

= 15PCA + SGK15.153+0.1470.1471.99534.26930.632
Kurtosis + SGK14.196+0.8040.8045.20833.89930.760
Mode + SGK12.494+2.5062.5060.08431.38147.311
Med + SGK14.838−0.1620.1620.08233.22130.971
Min + SGK1.807−13.19313.1930.02928.27496.766

= 20PCA + SGK20.071+0.0710.0711.61831.91141.876
Kurtosis + SGK18.943−1.0571.0574.81831.57445.256
Mode + SGK17.038−2.9632.9630.09430.18662.300
Med + SGK19.319−0.6810.6810.05531.77743.193
Min + SGK1.877−18.12318.1230.02822.152396.153

= 30PCA+SGK29.6720.3280.3281.39329.91766.286
Kurtosis + SGK28.480−1.5201.5205.12629.72469.292
Mode + SGK23.356−6.6446.6440.08429.044161.709
Med + SGK28.199−1.8011.8010.05429.61471.074
Min + SGK3.292−26.70826.7080.02918.675882.180

= 40PCA + SGK39.5460.4540.4541.86528.37194.630
Kurtosis + SGK38.238−1.7621.7625.13428.30596.062
Mode + SGK33.463−6.5376.5370.08725.813170.526
Med + SGK37.068−2.9322.9320.05327.910105.212
Min + SGK5.385−34.61534.6150.03416.2221152.049