Edit2Restore:通过参数高效适应预训练编辑模型实现少样本图像修复
Edit2Restore:Few-Shot Image Restoration via Parameter-Efficient Adaptation of Pre-trained Editing Models
AI总结：
Edit2Restore通过参数高效微调预训练编辑模型，在少样本条件下实现多种图像修复任务，提升感知质量并减少数据需求。
摘要：
图像修复传统上需要在每种退化类型上训练专门的模型，使用成千上万的配对示例。我们通过证明强大的预训练文本条件图像编辑模型可以通过参数高效微调来高效适应多种修复任务，从而挑战这一范式。我们的方法在FLUX.1 Kontext上微调LoRA适配器，这是一个最先进的12B参数流匹配模型，用于图像到图像转换。仅使用16-128个配对图像和简单的文本提示指导修复操作。与现有方法不同，现有方法从头开始训练专门的修复网络，使用成千上万个样本，我们利用大规模预训练编辑模型中已经编码的丰富视觉先验，大幅减少数据需求，同时保持高质量的感知质量。一个统一的LoRA适配器，通过任务特定的文本提示，可以有效处理多种退化，包括去噪、去雨和去雾。通过全面的消融研究，我们分析：（i）训练集大小对修复质量的影响，（ii）任务特定与统一多任务适配器之间的权衡，（iii）文本编码器微调的作用，以及（iv）零样本基线性能。虽然我们的方法优先考虑感知质量而不是像素完美重建指标如PSNR/SSIM，但我们的结果表明，预训练图像编辑模型在适当适应的情况下，提供了一种有竞争力且数据高效的替代传统图像修复方法，为少样本、提示引导的图像增强开辟了新途径。代码可在：https://github.com/makinyilmaz/Edit2Restore 复现我们的结果。
英文摘要：
Image restoration has traditionally required training specialized models on thousands of paired examples per degradation type. Large pre-trained text-conditioned image editing models encode rich priors about image structure, quality, and degradation, yet we find that this knowledge does not, on its own, make them restorers: state-of-the-art editing models largely fail at restoration in the zero-shot regime. We show that what these priors lack is not capability but direction, and that a small amount of parameter-efficient adaptation supplies it. Fine-tuning LoRA adapters on FLUX.1 Kontext, a 12B-parameter flow matching model for image-to-image translation, with only 32--128 paired images per task and guided by simple text prompts, we turn a mediocre zero-shot editor into a competitive restorer. A single unified adapter, conditioned on task-specific prompts, handles five diverse degradations. Despite using three to four orders of magnitude less data, our few-shot model surpasses a recent restoration baseline trained on over a million curated pairs on the majority of perceptual and distribution-level metrics, on which we evaluate in keeping with our focus on perceptual rather than pixel-fidelity quality. Through comprehensive studies, we analyze the impact of training-set size, the trade-off between task-specific and unified multi-task adapters, the effect of text encoder adaptation, and zero-shot baseline performance, establishing pre-trained editing models as a compelling, data-efficient foundation for few-shot, prompt-guided image restoration.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
