PSMP-CLIP:基于CLIP的零样本异常检测中的补丁提示SAM与多语义提示
PSMP-CLIP: Patch-Prompt SAM and Multi-Semantic Prompting for CLIP-Based Zero-Shot Anomaly Detection
- Harbin Engineering University(哈尔滨工程大学)
- Key Laboratory of Advanced Marine Communication and Information Technology(先进海洋通信与信息技术重点实验室)
- University of Toronto(多伦多大学)
- Kanagawa University(神奈川大学)
机构由 AI 辅助整理,请以论文原文为准。
AI总结:
针对现有CLIP方法异常图粗糙和提示有限的问题,提出PSMP-CLIP,结合补丁提示SAM2分割与多语义提示正则化,在14个数据集上取得领先的像素级AUROC。
AI中文摘要:
零样本异常检测旨在无需目标域样本的情况下定位异常。现有的基于CLIP的方法存在异常图粗糙和语义提示有限的问题。我们提出PSMP-CLIP,它集成了补丁提示SAM2分割(PPSS)和多语义引导的提示正则化(MSGPR)。PPSS直接从中间补丁特征中采样提示,避免了阈值漂移,并引导SAM2生成精确的掩码。MSGPR使用多个由语义锚点约束的可学习提示来保持泛化能力。在14个数据集上的实验显示出极具竞争力的性能,在MVTec AD、BTAD、DTD-Synthetic、CVC-ClinicDB、TN3K、Endo和Kvasir上取得了最佳的像素级AUROC。
英文摘要:
Zero-shot anomaly detection aims to localize anomalies without target-domain samples. Existing CLIP-based methods suffer from coarse anomaly maps and limited semantic prompts. We propose PSMP-CLIP, integrating patch-prompt SAM2 segmentation (PPSS) and multi-semantic guided prompt regularization (MSGPR). PPSS samples prompts directly from intermediate patch features, avoiding threshold drift and guiding SAM2 to produce precise masks. MSGPR uses multiple learnable prompts constrained by semantic anchors to preserve generalization. Experiments on 14 datasets show highly competitive performance, achieving the best pixel-level AUROC on MVTec AD, BTAD, DTD-Synthetic, CVC-ClinicDB, TN3K, Endo, and Kvasir.