用预训练扩散模型无需训练即可显著提升暗光图像质量
Zero-Shot Low Light Image Enhancement with Diffusion Prior
- 直接利用预训练扩散模型的内在特征指导图像增强
- 在多个数据集上超越现有最先进方法，颜色还原更准确
- 无需调参或微调，适用于低光增强和自动白平衡任务
本文提出一种简单高效的零样本暗光图像增强方法，旨在将暗光图像恢复为光照充足的视觉效果。该方法无需优化、训练、微调、文本条件或超参数调整，即可稳定实现高质量重建。我们利用大规模自然图像训练的文本到图像扩散模型的预训练先验，通过模型内部特征引导推理过程，不同于依赖定制约束的现有方法。定量评估显示，本方法在多个标准数据集上优于当前最优技术，定性分析表明颜色准确性提升，细微色偏得到修正。此外，无需任何修改，该方法在自动白平衡（AWB）任务中也达到与最先进方法相当的性能。
原文摘要 · Abstract (English)
In this paper, we present a simple yet highly effective "free lunch" solution for low-light image enhancement (LLIE), which aims to restore low-light images as if acquired in well-illuminated environments. Our method necessitates no optimization, training, fine-tuning, text conditioning, or hyperparameter adjustments, yet it consistently reconstructs low-light images with superior fidelity. Specifically, we leverage a pre-trained text-to-image diffusion prior, learned from training on a large collection of natural images, and the features present in the model itself to guide the inference, in contrast to existing methods that depend on customized constraints. Comprehensive quantitative evaluations demonstrate that our approach outperforms SOTA methods on established datasets, while qualitative analyses indicate enhanced color accuracy and the rectification of subtle chromatic deviations. Furthermore, additional experiments reveal that our method, without any modifications, achieves SOTA-comparable performance in the auto white balance (AWB) task.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。