提出高效对抗净化框架，实现快速高质图像修复。
DBLP: Noise Bridge Consistency Distillation For Efficient And Reliable Adversarial Purification
- 设计噪声桥蒸馏机制，对齐对抗噪声与干净数据分布。
- 达到顶尖鲁棒准确率，推理仅需约0.2秒，图像质量优异。
- 适合需要实时防御的部署场景，如安防、自动驾驶。
深度神经网络在众多任务中取得显著进展，但其对对抗扰动的敏感性仍是关键弱点。现有基于扩散模型的对抗净化方法通常需大量迭代去噪，严重限制实际应用。本文提出扩散桥蒸馏净化（DBLP）框架，核心是噪声桥蒸馏目标，在潜在一致性模型（LCM）中建立对抗噪声分布与干净数据分布之间的合理对齐。为提升语义保真度，引入自适应语义增强，将多尺度金字塔边缘图作为条件输入以引导净化过程。在多个数据集上的实验证明，DBLP实现了最先进的鲁棒准确率，图像质量优越，推理时间约为0.2秒，标志着向实时对抗净化迈出重要一步。
原文摘要 · Abstract (English)
Recent advances in deep neural networks (DNNs) have led to remarkable success across a wide range of tasks. However, their susceptibility to adversarial perturbations remains a critical vulnerability. Existing diffusion-based adversarial purification methods often require intensive iterative denoising, severely limiting their practical deployment. In this paper, we propose Diffusion Bridge Distillation for Purification (DBLP), a novel and efficient diffusion-based framework for adversarial purification. Central to our approach is a new objective, noise bridge distillation, which constructs a principled alignment between the adversarial noise distribution and the clean data distribution within a latent consistency model (LCM). To further enhance semantic fidelity, we introduce adaptive semantic enhancement, which fuses multi-scale pyramid edge maps as conditioning input to guide the purification process. Extensive experiments across multiple datasets demonstrate that DBLP achieves state-of-the-art (SOTA) robust accuracy, superior image quality, and around 0.2s inference time, marking a significant step toward real-time adversarial purification.
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。