CDGP：用于弱监督异常分割的对比双高斯过程
CDGP: Contrastive Dual Gaussian Processes for Weakly Supervised Anomaly Segmentation
浏览论文内容
中文总结 AI 辅助
针对工业视觉检测中像素级缺陷标注成本高的问题，提出弱监督框架CDGP，通过建模预测分布与后验优势统计量实现异常分割，在MVTec AD~2等数据集上定位指标表现优异。
中文摘要 AI 辅助
工业视觉检测需同时判断产品是否存在缺陷并定位缺陷，但大规模收集像素级标注的成本高昂。多数异常分割方法仅从无缺陷图像学习，并对偏离正常的区域打分。然而，真实缺陷与异常但正常的区域均可能大幅偏离正常，进而获得相似的高分。本文提出对比双高斯过程（Contrastive Dual Gaussian Processes，CDGP），这是一种弱监督框架，用于对密集 token 建模正常和异常诱导变量的预测分布。其后验优势统计量通过联合预测不确定性对预测均值差进行标准化，既提供空间证据又提供图像级置信度，该证据与分层正常重构残差互补以实现精细定位。所有校准仅使用训练数据，无需人工像素标注或测试时拟合。在 MVTec AD~2、KSDD2 和 VisA 数据集上，CDGP 在所有 MVTec AD~2 定位指标中排名第一，在 KSDD2 和 VisA 上也位居前列或具有竞争力。因子化和匹配的线性头控制明确了线性核高斯过程（Gaussian Process，GP）公式的贡献与范围。
英文摘要
Industrial visual inspection must both decide whether a product is defective and localize the defect, yet pixel-level masks are costly to collect at scale. Most anomaly-segmentation methods learn only from defect-free images and score deviations from normality. A true defect and an unusual-but-normal region, however, can both deviate substantially and receive similarly high scores. We propose Contrastive Dual Gaussian Processes (CDGP), a weakly supervised framework that models normal and anomaly inducing-variable predictive distributions over dense tokens. Its posterior-dominance statistic standardizes their predictive-mean difference by the joint predictive uncertainty, providing both spatial evidence and image-level confidence. This evidence complements hierarchical normal-reconstruction residuals for fine localization. All calibration uses training data only, without human pixel annotations or test-time fitting. Across MVTec AD~2, KSDD2, and VisA, CDGP ranks first among the evaluated methods on all MVTec AD~2 localization metrics and is first-place or competitive on KSDD2 and VisA. Factorized and matched linear-head controls delimit the contribution and scope of the linear-kernel Gaussian process (GP) formulation.
发表机构
- Korea University（高丽大学）
- Dartmouth College（达特茅斯学院）
机构由 AI 辅助整理，请以论文原文为准。
