EGSDE: Unpaired Image-to-Image Translation via Energy-Guided Stochastic Differential Equations

TL;DR

EGSDE introduces an energy-guided approach using pretrained energy functions to enhance unpaired image-to-image translation, improving realism and faithfulness over existing diffusion-based methods.

Contribution

The paper proposes EGSDE, a novel energy-guided stochastic differential equation framework that leverages energy functions trained on both domains to improve unpaired I2I translation.

Findings

EGSDE outperforms baseline methods on multiple I2I tasks.

Achieves state-of-the-art realism metrics without sacrificing faithfulness.

Enables flexible trade-offs between realism and faithfulness.

Abstract

Score-based diffusion models (SBDMs) have achieved the SOTA FID results in unpaired image-to-image translation (I2I). However, we notice that existing methods totally ignore the training data in the source domain, leading to sub-optimal solutions for unpaired I2I. To this end, we propose energy-guided stochastic differential equations (EGSDE) that employs an energy function pretrained on both the source and target domains to guide the inference process of a pretrained SDE for realistic and faithful unpaired I2I. Building upon two feature extractors, we carefully design the energy function such that it encourages the transferred image to preserve the domain-independent features and discard domain-specific ones. Further, we provide an alternative explanation of the EGSDE as a product of experts, where each of the three experts (corresponding to the SDE and two feature extractors) solely contributes to faithfulness or realism. Empirically, we compare EGSDE to a large family of baselines on three widely-adopted unpaired I2I tasks under four metrics. EGSDE not only consistently outperforms existing SBDMs-based methods in almost all settings but also achieves the SOTA realism results without harming the faithful performance. Furthermore, EGSDE allows for flexible trade-offs between realism and faithfulness and we improve the realism results further (e.g., FID of 51.04 in Cat to Dog and FID of 50.43 in Wild to Dog on AFHQ) by tuning hyper-parameters. The code is available at https://github.com/ML-GSAI/EGSDE.