Guided Image-to-Image Translation with Bi-Directional Feature Transformation

ICCV 2019  ยท  Badour AlBahar, Jia-Bin Huang ยท

We address the problem of guided image-to-image translation where we translate an input image into another while respecting the constraints provided by an external, user-provided guidance image. Various conditioning methods for leveraging the given guidance image have been explored, including input concatenation , feature concatenation, and conditional affine transformation of feature activations. All these conditioning mechanisms, however, are uni-directional, i.e., no information flow from the input image back to the guidance. To better utilize the constraints of the guidance image, we present a bi-directional feature transformation (bFT) scheme. We show that our bFT scheme outperforms other conditioning schemes and has comparable results to state-of-the-art methods on different tasks.

PDF Abstract ICCV 2019 PDF ICCV 2019 Abstract

Datasets


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Pose Transfer Deep-Fashion bFT SSIM 0.767 # 7
IS 3.22 # 6
FID 12.266 # 2
Image Reconstruction Edge-to-Clothes bFT FID 58.4 # 1
LPIPS 0.1 # 1
Image Reconstruction Edge-to-Handbags bFT FID 74.9 # 3
LPIPS 0.2 # 2
Image Reconstruction Edge-to-Shoes bFT FID 121.2 # 3
LPIPS 0.1 # 3

Methods


No methods listed for this paper. Add relevant methods here