Image-to-Image Translation with Conditional Adversarial Networks

CVPR 2017 Phillip IsolaJun-Yan ZhuTinghui ZhouAlexei A. Efros

We investigate conditional adversarial networks as a general-purpose solution to image-to-image translation problems. These networks not only learn the mapping from input image to output image, but also learn a loss function to train this mapping... (read more)

PDF Abstract

Code


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT LEADERBOARD
Image-to-Image Translation Aerial-to-Map cGAN Per-pixel Accuracy 70% # 1
Per-class Accuracy 46% # 1
Class IOU 0.26 # 1
Facial Expression Translation AR Face Enc.-Decoder AMT 0.1 # 2
PSNR 12.6660 # 2
Nuclear Segmentation Cell17 Pix2Pix F1-score 0.6208 # 5
Dice 0.6351 # 4
Hausdorff 19.1441 # 4
Image-to-Image Translation Cityscapes Labels-to-Photo pix2pix Class IOU 0.18 # 1
Per-class Accuracy 25% # 1
Per-pixel Accuracy 71% # 6
Image-to-Image Translation Cityscapes Photo-to-Labels pix2pix Per-pixel Accuracy 85% # 1
Per-class Accuracy 40% # 1
Class IOU 0.32 # 1
Cross-View Image-to-Image Translation cvusa Pix2pix SSIM 0.3923 # 7
Cross-View Image-to-Image Translation Dayton (256×256) - aerial-to-ground Pix2pix SSIM 0.4180 # 5
Cross-View Image-to-Image Translation Dayton (256×256) - ground-to-aerial Pix2pix SSIM 0.2693 # 4
Cross-View Image-to-Image Translation Dayton (64×64) - aerial-to-ground Pix2pix SSIM 0.4808 # 5
Cross-View Image-to-Image Translation Dayton (64x64) - ground-to-aerial Pix2pix SSIM 0.3675 # 3
Cross-View Image-to-Image Translation Ego2Top Pix2pix SSIM 0.2213 # 4

Results from Other Papers


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK SOURCE PAPER COMPARE
Image Reconstruction Edge-to-Handbags pix2pix FID 96.31 # 4
LPIPS 0.234 # 4
Image Reconstruction Edge-to-Shoes pix2pix FID 197.492 # 4
LPIPS 0.238 # 1