Unpaired Image-to-Image Translation via a Self-Supervised Semantic Bridge

TL;DR

This paper introduces the Self-Supervised Semantic Bridge (SSB), a novel framework that uses semantic priors and self-supervised encoders to improve unpaired image translation, especially in medical imaging and text-guided editing.

Contribution

The paper presents SSB, a new method that integrates semantic priors into diffusion models, enabling spatially faithful translation without cross-domain supervision.

Findings

SSB outperforms existing methods in medical image synthesis.

SSB achieves high-quality, out-of-domain translation.

SSB extends to text-guided image editing.

Abstract

Adversarial diffusion and diffusion-inversion methods have advanced unpaired image-to-image translation, but each faces key limitations. Adversarial approaches require target-domain adversarial loss during training, which can limit generalization to unseen data, while diffusion-inversion methods often produce low-fidelity translations due to imperfect inversion into noise-latent representations. In this work, we propose the Self-Supervised Semantic Bridge (SSB), a versatile framework that integrates external semantic priors into diffusion bridge models to enable spatially faithful translation without cross-domain supervision. Our key idea is to leverage self-supervised visual encoders to learn representations that are invariant to appearance changes but capture geometric structure, forming a shared latent space that conditions the diffusion bridges. Extensive experiments show that SSB outperforms strong prior methods for challenging medical image synthesis in both in-domain and out-of-domain settings, and extends easily to high-quality text-guided editing.