Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks

ICCV 2017 Jun-Yan ZhuTaesung ParkPhillip IsolaAlexei A. Efros

Image-to-image translation is a class of vision and graphics problems where the goal is to learn the mapping between an input image and an output image using a training set of aligned image pairs. However, for many tasks, paired training data will not be available... (read more)

PDF Abstract

Code


Evaluation results from the paper


Task Dataset Model Metric name Metric value Global rank Compare
Multimodal Unsupervised Image-To-Image Translation Cats-and-Dogs CycleGAN CIS 0.076 # 3
Multimodal Unsupervised Image-To-Image Translation Cats-and-Dogs CycleGAN IS 0.813 # 3
Image-to-Image Translation Cityscapes Labels-to-Photo CycleGAN Class IOU 0.11 # 2
Image-to-Image Translation Cityscapes Labels-to-Photo CycleGAN Per-class Accuracy 17% # 2
Image-to-Image Translation Cityscapes Labels-to-Photo CycleGAN Per-pixel Accuracy 52% # 6
Image-to-Image Translation Cityscapes Photo-to-Labels CycleGAN Per-pixel Accuracy 58% # 2
Image-to-Image Translation Cityscapes Photo-to-Labels CycleGAN Per-class Accuracy 22% # 2
Image-to-Image Translation Cityscapes Photo-to-Labels CycleGAN Class IOU 0.16 # 2
Multimodal Unsupervised Image-To-Image Translation Edge-to-Handbags CycleGAN Quality 40.8% # 3
Multimodal Unsupervised Image-To-Image Translation Edge-to-Handbags CycleGAN Diversity 0.012 # 4
Multimodal Unsupervised Image-To-Image Translation Edge-to-Shoes CycleGAN Quality 36.0% # 4
Multimodal Unsupervised Image-To-Image Translation Edge-to-Shoes CycleGAN Diversity 0.010 # 4
Image-to-Image Translation RaFD CycleGAN Classification Error 5.99% # 3