SPatchGAN: A Statistical Feature Based Discriminator for Unsupervised Image-to-Image Translation

TL;DR

SPatchGAN introduces a statistical feature-based discriminator for unsupervised image-to-image translation, improving stability and detail preservation without complex constraints, outperforming state-of-the-art models across diverse tasks.

Contribution

The paper presents a novel discriminator architecture that emphasizes statistical features over patches, simplifying the framework and enhancing translation quality.

Findings

Outperforms existing models in selfie-to-anime translation

Effective in male-to-female and glasses removal tasks

Stabilizes training through multi-scale distribution matching

Abstract

For unsupervised image-to-image translation, we propose a discriminator architecture which focuses on the statistical features instead of individual patches. The network is stabilized by distribution matching of key statistical features at multiple scales. Unlike the existing methods which impose more and more constraints on the generator, our method facilitates the shape deformation and enhances the fine details with a greatly simplified framework. We show that the proposed method outperforms the existing state-of-the-art models in various challenging applications including selfie-to-anime, male-to-female and glasses removal.