用扩散模型修复损坏感知数据，提升多车协同的鲁棒性
Diff-KD: Diffusion-based Knowledge Distillation for Collaborative Perception under Corruptions
- 将特征恢复建模为条件扩散过程，主动重建被干扰的语义
- 在7种干扰下，检测准确率与校准鲁棒性均达当前最优
- 适合自动驾驶多车协同感知场景，尤其应对传感器/通信异常
多智能体协同感知通过集体智能克服个体传感局限，但现实中的传感器与通信干扰严重削弱这一优势。现有方法将干扰视为静态扰动或被动适应，无法主动恢复原始清洁语义。为此，我们提出Diff-KD框架，将基于扩散的生成修复融入师生知识蒸馏，以实现鲁棒的协同感知。该框架包含两个核心组件：(i) 渐进式知识蒸馏（PKD），将局部特征恢复建模为条件扩散过程，从受损观测中恢复全局语义；(ii) 自适应门控融合（AGF），根据自身可靠性动态加权邻居信息。在OPV2V和DAIR-V2X数据集上，针对七类干扰类型进行评估，Diff-KD在检测准确率与校准鲁棒性方面均达到当前最优水平。
原文摘要 · Abstract (English)
Multi-agent collaborative perception enables autonomous systems to overcome individual sensing limits through collective intelligence. However, real-world sensor and communication corruptions severely undermine this advantage. Crucially, existing approaches treat corruptions as static perturbations or passively conform to corrupted inputs, failing to actively recover the underlying clean semantics. To address this limitation, we introduce Diff-KD, a framework that integrates diffusion-based generative refinement into teacher-student knowledge distillation for robust collaborative perception. Diff-KD features two core components: (i) Progressive Knowledge Distillation (PKD), which treats local feature restoration as a conditional diffusion process to recover global semantics from corrupted observations; and (ii) Adaptive Gated Fusion (AGF), which dynamically weights neighbors based on ego reliability during fusion. Evaluated on OPV2V and DAIR-V2X under seven corruption types, Diff-KD achieves state-of-the-art performance in both detection accuracy and calibration robustness.
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。