Images from the MIT-Adobe 5K dataset

First column shows the input images.

Second column shows the retouched images by Expert E.

Third column shows the predicted results for HDRNet.

Fourth and fifth column shows the results for 3 dimensions using PCA and VAE respectively.

Finally we show the results using our method in the last column.

You can toggle between the predicted and the target image by clicking on the images. Moreover, you can rearrange the images in increasing or decreasing order of the metrics.

ID Input Expert E HDRNet PSNR 21.057 FLIP 0.363 Delta E 2000 10.779 PCA PSNR 17.681 FLIP 0.421 Delta E 2000 12.461 VAE PSNR 27.838 FLIP 0.183 Delta E 2000 4.290 Ours PSNR 36.352 FLIP 0.120 Delta E 2000 2.366
000001 24.704 0.299 6.113 22.067 0.378 9.976 26.322 0.192 4.621 36.061 0.125 1.905
000002 27.841 0.192 5.853 16.738 0.450 11.019 33.394 0.128 3.108 35.212 0.105 2.165
000003 20.701 0.367 9.190 7.029 0.801 38.344 23.446 0.246 6.463 25.721 0.277 5.136
000004 17.806 0.431 10.806 23.947 0.214 5.610 28.772 0.145 3.529 47.623 0.044 0.731
000005 16.184 0.453 14.909 17.779 0.597 17.498 27.299 0.181 3.388 36.278 0.115 2.011
000006 18.285 0.423 14.211 19.369 0.408 10.257 21.338 0.325 6.925 34.255 0.122 2.544
000007 26.500 0.211 6.963 15.160 0.516 14.201 28.171 0.155 3.710 30.275 0.164 2.755
000008 18.436 0.477 13.672 19.911 0.447 12.031 24.445 0.225 4.623 29.562 0.193 3.623
000009 21.829 0.366 7.805 23.485 0.277 6.647 29.226 0.155 2.836 41.508 0.078 1.555
000010 26.235 0.219 4.963 21.825 0.222 6.087 28.552 0.163 3.963 51.811 0.038 1.049