通过解耦表征与动态融合，提升视觉语言模型在零样本异常检测中的精度与鲁棒性。
ACD-CLIP: Decoupling Representation and Dynamic Fusion for Zero-Shot Anomaly Detection
- 采用卷积低秩适配器注入局部先验知识，增强细粒度表征能力
- 设计动态融合网关，根据视觉上下文自适应调制文本提示，实现双向信息融合
- 在工业与医疗多场景数据集上表现优异，适合需要高可靠性异常检测的应用
预训练视觉语言模型在零样本异常检测任务中表现受限，主要源于两个关键问题：缺乏密集预测所需的局部归纳偏置，以及固定不变的特征融合范式。为此，本文提出一种架构协同设计框架，同步优化特征表征与跨模态融合。方法引入参数高效的卷积低秩适配器（Conv-LoRA），以注入局部归纳偏置，提升细粒度表征能力；并设计动态融合网关（DFG），利用视觉上下文自适应调制文本提示，实现强大的双向融合。在多个工业与医疗基准数据集上的大量实验表明，该方法显著提升了检测精度与鲁棒性，验证了协同设计对基础模型向密集感知任务迁移的关键作用。代码已开源：https://github.com/cockmake/ACD-CLIP。
原文摘要 · Abstract (English)
Pre-trained Vision-Language Models (VLMs) struggle with Zero-Shot Anomaly Detection (ZSAD) due to a critical adaptation gap: they lack the local inductive biases required for dense prediction and employ inflexible feature fusion paradigms. We address these limitations through an Architectural Co-Design framework that jointly refines feature representation and cross-modal fusion. Our method proposes a parameter-efficient Convolutional Low-Rank Adaptation (Conv-LoRA) adapter to inject local inductive biases for fine-grained representation, and introduces a Dynamic Fusion Gateway (DFG) that leverages visual context to adaptively modulate text prompts, enabling a powerful bidirectional fusion. Extensive experiments on diverse industrial and medical benchmarks demonstrate superior accuracy and robustness, validating that this synergistic co-design is critical for robustly adapting foundation models to dense perception tasks. The source code is available at https://github.com/cockmake/ACD-CLIP.
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。