RefGC-SR$^2$: Reference-guided Generated Content Super-Resolution and Refinement

2026-06-13

AI Summary

This work addresses critical limitations in existing reference-guided generation and restoration methods, which either lose fine details due to downsampling high-resolution references during input or introduce identity distortions and other artifacts during generation. Furthermore, current inpainting and super-resolution approaches are often confined to low-resolution domains or fail to account for the unique characteristics of generative artifacts. To overcome these issues, we propose a novel task, RefGC-SR², which leverages the original high-resolution reference image in a post-processing stage to simultaneously restore fine details, correct artifacts, and perform image upscaling. We introduce the first real-world triplet data synthesis pipeline tailored to this task and develop a frequency-aware diffusion Transformer that enables selective injection of reference details while jointly suppressing artifacts. Experiments demonstrate that our method substantially outperforms existing RefGCR and RefSR baselines, significantly enhancing visual quality and practical utility while preserving identity consistency.

Abstract

Reference-guided generation (e.g., object compositing, customization) has progressed rapidly, yet current pipelines share a fundamental limitation: the object-centric high-resolution reference image (HRRI) provided by users is downsampled to a fixed low-resolution (LR) before being fed into the model, so the fine-grained details are discarded before the output is even produced. In addition, the generation step then introduces its own artifacts (e.g., identity distortion) on top of this loss. Existing reference-guided generated content refinement (RefGCR) methods can correct some of these artifacts but still operate in the LR domain; reference-guided super-resolution (RefSR) methods recover resolution but assume natural-image degradations and ignore the artifact distribution of generative pipelines. To address both gaps in a single formulation, we introduce a new task: reference-guided generated content super-resolution-refinement (RefGC-SR$^2$), where the original HRRI is reused at the post-processing stage to recover lost details, refine generative artifacts, and upscale the output simultaneously. We construct the first real-world triplet data generation pipeline for this RefGC-SR$^2$ task, training a diptych-conditioned generator to synthesize paired low-quality anchors that public pretrained models cannot provide. We further present a frequency-aware diffusion transformer model for RefGC-SR$^2$ that selectively injects fine details from the HRRI while removing generative artifacts. Extensive experiments demonstrate that our RefGC-SR$^2$ model successfully (i) refines the object identity faithfully with respect to the reference, and (ii) recovers high-resolution details, so that the final result is significantly higher quality and practically more usable compared to existing RefGCR and RefSR baselines.

Problem

• reference-guided generation
• super-resolution
• generative artifacts
• detail recovery
• image refinement

Innovation

• reference-guided generation
• super-resolution
• artifact refinement
• diffusion transformer
• frequency-aware