Unsupervised Structure-Consistent Image-to-Image Translation

TL;DR

This paper introduces a new unsupervised image-to-image translation method that enhances structure-texture disentanglement and control, reduces training time, and performs well on complex datasets like satellite images.

Contribution

It proposes an auxiliary module with gradient reversal layers for better disentanglement and structure consistency without semantic masks, improving image translation quality and efficiency.

Findings

Outperforms state-of-the-art in complex domains like satellite images.

Reduces training time significantly.

Improves quality metrics across various datasets.

Abstract

The Swapping Autoencoder achieved state-of-the-art performance in deep image manipulation and image-to-image translation. We improve this work by introducing a simple yet effective auxiliary module based on gradient reversal layers. The auxiliary module's loss forces the generator to learn to reconstruct an image with an all-zero texture code, encouraging better disentanglement between the structure and texture information. The proposed attribute-based transfer method enables refined control in style transfer while preserving structural information without using a semantic mask. To manipulate an image, we encode both the geometry of the objects and the general style of the input images into two latent codes with an additional constraint that enforces structure consistency. Moreover, due to the auxiliary loss, training time is significantly reduced. The superiority of the proposed model is demonstrated in complex domains such as satellite images where state-of-the-art are known to fail. Lastly, we show that our model improves the quality metrics for a wide range of datasets while achieving comparable results with multi-modal image generation techniques.