Skip to main content

Table 3 Image denoising on depth images with NLM ,BM3D, OGLR, ADNet and our method: peformance comparisons in PSNR (Left, in dB) and SSIM (Right)

From: Optimal graph edge weights driven nlms with multi-layer residual compensation

Images noise NLM BM3D OGLR ADNet Our Method
teddy \({\sigma }=10\) 37.04dB 0.9725 40.10dB 0.9818 39.98dB 0.9833 41.98dB 0.9878 38.24dB 0.9673
\({\sigma }=20\) 33.60dB 0.9336 35.94dB 0.9674 35.94dB 0.9644 37.74dB 0.9779 34.69dB 0.9644
\({\sigma }=30\) 30.89dB 0.8842 33.16dB 0.9481 33.49dB 0.9441 35.07dB 0.9669 33.02dB 0.9486
\({\sigma }=40\) 29.04dB 0.8299 31.32dB 0.9279 31.78dB 0.9277 33.15dB 0.9539 30.91dB 0.9288
\({\sigma }=50\) 27.71dB 0.7741 29.73dB 0.9190 30.46dB 0.9049 31.50dB 0.9404 30.38dB 0.9125
wood \({\sigma }=10\) 43.64dB 0.9870 42.21dB 0.9889 44.45dB 0.9882 43.46dB 0.9924 42.50dB 0.9701
\({\sigma }=20\) 38.39dB 0.9559 38.09dB 0.9727 40.56dB 0.9761 39.79dB 0.9862 41.09dB 0.9839
\({\sigma }=30\) 35.17dB 0.9137 35.94dB 0.9573 38.15dB 0.9600 37.82dB 0.9792 38.91dB 0.9770
\({\sigma }=40\) 32.83dB 0.8620 34.46dB 0.9420 36.57dB 0.9478 36.58dB 0.9746 37.47dB 0.9656
\({\sigma }=50\) 31.06dB 0.8159 33.28dB 0.9375 34.55dB 0.9184 35.41dB 0.9664 36.07dB 0.9512
Sawtooth \({\sigma }=10\) 43.77dB 0.9866 44.62dB 0.9884 45.35dB 0.9911 45.61dB 0.9912 43.40dB 0.9731
\({\sigma }=20\) 38.12dB 0.9513 41.16dB 0.9787 40.96dB 0.9779 42.72dB 0.9873 41.11dB 0.9865
\({\sigma }=30\) 34.78dB 0.9032 38.72dB 0.9655 38.34dB 0.9630 40.41dB 0.9802 39.56dB 0.9762
\({\sigma }=40\) 32.69dB 0.8484 36.84dB 0.9497 36.51dB 0.9526 38.84dB 0.9747 37.29dB 0.9665
\({\sigma }=50\) 31.19dB 0.7912 35.87dB 0.9500 34.94dB 0.9334 37.34dB 0.9654 35.84dB 0.9510
Tsukuba \({\sigma }=10\) 39.71dB 0.9724 41.52dB 0.9798 41.64dB 0.9850 41.41dB 0.9813 40.67dB 0.9701
\({\sigma }=20\) 35.53dB 0.9308 37.38dB 0.9577 37.56dB 0.9626 37.76dB 0.9667 37.28dB 0.9657
\({\sigma }=30\) 32.37dB 0.8743 34.81dB 0.9319 34.91dB 0.9403 35.58dB 0.9486 35.23dB 0.9468
\({\sigma }=40\) 30.05dB 0.8110 33.05dB 0.9053 32.99dB 0.9245 34.00dB 0.9244 33.41dB 0.9313
\({\sigma }=50\) 28.36dB 0.7497 31.67dB 0.8915 31.44dB 0.9003 32.78dB 0.9157 32.57dB 0.9133
Books \({\sigma }=10\) 40.28dB 0.9727 40.89dB 0.9811 42.37dB 0.9828 41.83dB 0.9838 41.69dB 0.9827
\({\sigma }=20\) 36.37dB 0.9351 35.10dB 0.9519 38.26dB 0.9625 37.01dB 0.9609 37.35dB 0.9666
\({\sigma }=30\) 33.89dB 0.8887 32.61dB 0.9235 35.88dB 0.9420 34.45dB 0.9461 35.62dB 0.9491
\({\sigma }=40\) 32.18dB 0.8373 31.22dB 0.8970 34.10dB 0.9256 32.44dB 0.9260 34.28dB 0.9334
\({\sigma }=50\) 30.86dB 0.7834 29.76dB 0.8812 32.86dB 0.9050 31.11dB 0.9147 33.31dB 0.9154
Bowling \({\sigma }=10\) 43.18dB 0.9850 41.62dB 0.9850 42.70dB 0.9867 43.09dB 0.9899 42.12dB 0.9725
\({\sigma }=20\) 38.65dB 0.9542 37.72dB 0.9682 38.84dB 0.9723 39.48dB 0.9796 39.32dB 0.9816
\({\sigma }=30\) 35.62dB 0.9126 35.30dB 0.9484 36.45dB 0.9573 37.46dB 0.9699 38.22dB 0.9732
\({\sigma }=40\) 33.32dB 0.8622 33.60dB 0.9379 34.73dB 0.9481 36.03dB 0.9624 36.27dB 0.9654
\({\sigma }=50\) 31.54dB 0.8069 32.18dB 0.9193 33.70dB 0.9330 34.37dB 0.9466 35.56dB 0.9531
  1. Bold and underline to mark the best and the second best results for each quality index, respectively