Skip to content

Benchmark (from papers)

Tang, Wenyi edited this page May 10, 2019 · 1 revision

Benchmark

NOTE: all results are copied from original papers.

SISR (Single-Image Super-Resolution)

Model Scale Factor Set5 Set14 BSD100 Urban100
Bicubic x2 33.66/0.9299 30.24/0.8688 29.56/0.8431 26.88/0.8403
SRCNN x2 36.66/0.9542 32.45/0.9067 31.36/0.8879 29.50/0.8946
ESPCN x2 -/- -/- -/- -/-
VDSR x2 37.53/0.9587 33.03/0.9124 31.90/0.8960 30.76/0.9140
DRCN x2 37.63/0.9588 33.04/0.9118 31.85/0.8942 30.75/0.9133
DRRN x2 37.74/0.9591 33.23/0.9136 32.05/0.8973 31.23/0.9188
LapSrn x2 37.52/0.959 33.08/0.913 31.80/0.895 30.41/0.910
EDSR x2 38.11/0.9601 33.92/0.9195 32.32/0.9013 32.93/0.9351
SRGAN x2 -/- -/- -/- -/-
MemNet x2 37.78/0.9597 33.28/0.9142 32.08/0.8978 31.31/0.9195
SRDense x2 -/- -/- -/- -/-
IDN x2 37.83/0.9600 33.30/0.9148 32.08/0.8985 31.27/0.9196
RDN x2 38.24/0.9614 34.01/0.9212 32.34/0.9017 32.89/0.9353
SRMD x2 37.53/0.959 33.12/0.914 31.90/0.896 30.89/0.916
SRMD(NF) x2 37.79/0.9601 33.32/0.9159 32.05/0.8985 31.33/0.9204
D-DBPN x2 38.09/0.9600 33.85/0.9190 32.27/0.9000 32.55/0.9324
CARN x2 37.76/0.9590 33.52/0.9166 32.09/0.8978 31.51/0.9312
RCAN x2 38.27/0.9614 34.12/0.9216 32.41/0.9027 33.34/0.9384
MSRN x2 38.08/0.9605 33.74/0.9170 32.23/0.9013 32.22/0.9326
SRFeat x2 -/- -/- -/- -/-
NLRN x2 38.00/0.9603 33.46/0.9159 32.19/0.8992 31.81/0.9249
RNAN x2 38.17/0.9611 33.87/0.9207 32.32/0.9014 32.73/0.9340
-------- ------------- ------------ ------------ ------------ ------------
Bicubic x4 28.42/0.8104 26.00/0.7027 25.96/0.6675 23.14/0.6577
SRCNN x4 30.48/0.8628 27.50/0.7513 26.90/0.7101 24.52/0.7221
ESPCN x4 30.90/- 27.83/- -/- -/-
VDSR x4 31.35/0.8838 28.01/0.7674 27.29/0.7251 25.18/0.7524
DRCN x4 31.53/0.8854 28.02/0.7670 27.23/0.7233 25.14/0.7510
DRRN x4 31.68/0.8888 28.21/0.7721 27.38/0.7284 25.44/0.7638
LapSrn x4 31.54/0.885 28.19/0.772 27.32/0.728 25.21/0.756
EDSR x4 32.46/0.8968 28.80/0.7876 27.71/0.7420 26.64/0.8033
SRGAN x4 29.40/0.8472 26.02/0.7397 25.16/0.6688 -/-
MemNet x4 31.74/0.8893 28.26/0.7723 27.40/0.7281 25.50/0.7630
SRDense x4 32.02/0.8934 28.50/0.7782 27.53/0.7337 26.05/0.7819
IDN x4 31.82/0.8903 28.25/0.7730 27.41/0.7297 25.41/0.7632
RDN x4 32.47/0.8990 28.81/0.7871 27.72/0.7419 26.61/0.8028
SRMD x4 31.59/0.887 28.15/0.772 27.34/0.728 25.34/0.761
SRMD(NF) x4 31.96/0.8925 28.35/0.7787 27.49/0.7337 25.68/0.7731
D-DBPN x4 32.47/0.8980 28.82/0.7860 27.72/0.7400 26.38/0.7946
CARN x4 32.13/0.8937 28.60/0.7806 27.58/0.7349 26.07/0.7837
RCAN x4 32.63/0.9002 28.87/0.7889 27.77/0.7436 26.82/0.8087
MSRN x4 32.07/0.8903 28.60/0.7751 27.52/0.7273 26.04/0.7896
SRFeat x4 32.27/0.8938 28.71/0.7835 27.64/0.7378 -/-
NLRN x4 31.92/0.8916 28.36/0.7745 27.48/0.7306 25.79/0.7729
RNAN x4 32.49/0.8982 28.83/0.7878 27.72/0.7421 26.61/0.8023
ESRGAN x4 32.74/0.9012 28.99/0.7917 27.85/0.7455 27.03/0.8153
-------- ------------- ------------ ------------ ------------ ------------
  • PSNR Top3: 1st ESRGAN, 2nd RCAN, 3rd RNAN, 4th RDN.
  • SSIM Top3: 1st ESRGAN, 2nd RCAN, 3rd RDN, 4th RNAN.

Image Denoise

Model AWGN Level BSD68 Set12 Urban100
BM3D 15 31.08/0.8722 32.37/0.8952 32.35/0.9220
TNRD 15 31.42/0.8826 32.50/0.8958 31.86/0.9031
DnCNN 15 31.46/0.8826 32.86/0.9031 32.68/0.9255
FFDNet 15 31.63/0.8902 32.75/0.9027
NLRN 15 31.88/0.8932 33.16/0.9070 33.45/0.9354
C-FFDNet 15 33.87/0.9288 -/-
C-DnCNN 15 33.87/0.9288 -/-
-------- ------------ ------------ ------------ ------------
BM3D 25 28.57/0.8013 29.97/0.8504 29.70/0.8777
TNRD 25 28.92/0.8093 30.06/0.8512 29.25/0.8473
DnCNN 25 29.23/0.8278 30.44/0.8622 29.97/0.8797
FFDNet 25 29.19/0.8289 30.43/0.8634
NLRN 25 29.41/0.8331 30.80/0.8689 30.94/0.9018
C-FFDNet 25 31.21/0.8817 -/-
C-DnCNN 25 31.21/0.8817 -/-
-------- ------------ ------------ ------------ ------------
BM3D 50 25.62/0.6864 26.72/0.7676 25.95/0.7791
TNRD 50 25.97/0.6994 26.81/0.7680 25.88/0.7563
DnCNN 50 26.23/0.7189 27.18/0.7829 26.28/0.7874
FFDNet 50 26.29/0.7245 27.32/0.7903
NLRN 50 26.47/0.7298 27.64/0.7980 27.49/0.8279
C-FFDNet 50 27.96/0.7881 -/-
C-DnCNN 50 27.96/0.7881 -/-
-------- ------------ ------------ ------------ ------------

NOTE C-***** means Color image denoising model.

Degradation Image Super-Resolution

PSNR on RGB channel.

Model Scale AWGN Set5 Set14 Urban100 BSD100
BM3D+Bi x4 15 24.52/0.6680 22.92/0.5747 20.93/0.5600 23.28/0.5556
SRMD x4 15 26.57/0.7530 24.31/0.6410 22.54/0.6538 24.46/0.6183
CBD+ESR x4 15 25.61/0.7187 23.74/0.6154 22.01/0.6275 24.10/0.6001
CBD+CARN x4 15 25.61/0.7198 23.70/0.6152 21.94/0.6250 24.09/0.5999
FFD+CARN x4 15 26.31/0.7432 24.21/0.6361 22.43/0.6477 24.38/0.6154
FFD+ESR x4 15 26.30/0.7421 22.53/0.6521 24.39/0.6159
-------- ------- ------ ------------ ------------ ------------ ------------
BM3D+Bi x4 25 23.24/0.6155 22.06/0.5348 20.33/0.5254 22.47/0.5194
SRMD x4 25 25.02/0.6999 23.23/0.5923 21.78/0.6124 23.60/0.5765
CBD+ESR x4 25 24.34/0.6728 22.69/0.5690 21.18/0.5817 23.24/0.5602
CBD+CARN x4 25 24.31/0.6730 22.68/0.5695 21.13/0.5805 23.24/0.5604
FFD+CARN x4 25 24.79/0.6883 23.20/0.5894 21.60/0.6026 23.51/0.5729
FFD+ESR x4 25 24.78/0.6872 21.65/0.6043 23.51/0.5731
-------- ------- ------ ------------ ------------ ------------ ------------
BM3D+Bi x4 50 20.54/0.5189 20.22/0.4716 18.90/0.4613 20.85/0.4635
SRMD x4 50 22.04/0.5996 21.10/0.5151 20.12/0.5345 21.85/0.5066
CBD+ESR x4 50 21.85/0.5763 20.54/0.4857 19.59/0.4964 21.63/0.4899
CBD+CARN x4 50 21.84/0.5770 20.55/0.4870 19.58/0.4967 21.64/0.4910
FFD+CARN x4 50 21.84/0.5877 21.23/0.5145 19.91/0.5216 21.76/0.5013
FFD+ESR x4 50 21.86/0.5877 19.93/0.5220 21.77/0.5014
-------- ------- ------ ------------ ------------ ------------ ------------

P.S: BM3D+Bicubic, CBDNet+ESRGAN_PSNR, FFDNet+CARN

Pretrained Models

VSR (Tensorflow)

  1. SRCNN

VSRTorch

  1. Classic
  2. CARN
  3. ESRGAN
  4. TecoGAN
  5. SOF-VSR
  6. FRVSR
  7. RBPN
Clone this wiki locally