DuET: Dual Expert Trajectories for Diffusion Image Editing
For users of diffusion-based image editing, DuET offers a training-free method to enhance edit fidelity and naturalness, particularly when target scenes diverge from the source.
DuET improves diffusion-based image editing by temporarily relaxing source-image conditioning, achieving better instruction relevance and perceptual quality without extra training or sampling cost, though with a slight trade-off in source preservation.
Recent diffusion editors perform diverse instruction-based edits while conditioning on the source image at every denoising step. Yet persistent source-image conditioning can limit how fully an edit is executed and how natural the result appears, especially when the target scene diverges substantially from the input. We introduce DuET (Dual Expert Trajectories), a training-free inference method that temporarily relaxes source-image conditioning by transitioning through a text-to-image phase before returning to edit mode, allowing the denoising trajectory to move toward the target distribution while retaining the structural benefits of image-conditioned editing. Without modifying model weights or increasing sampling cost, DuET consistently improves instruction relevance, semantic fidelity, and perceptual quality across diverse models and benchmarks. In some cases, these gains come with a modest reduction in source-image preservation, revealing a predictable trade-off between source preservation and edit fidelity.