提出新方法评估工业缺陷检测的误报率与空间证据，更真实反映模型性能。
Distribution-free false-alarm calibration and chance-corrected spatial evaluation for industrial anomaly detection
- 用无分布上界阈值和配对交叉测试，校准误报率并扣除偶然重合影响。
- 实测显示仅DINOv2-ASM在多数场景有显著空间证据提升（最高0.347）。
- 适合关注实际部署性能、避免误判误导的研究者和工程师。
工业视觉检测研究常报告受试者工作特征曲线下面积（AUROC）及异常图与缺陷掩码的重叠度，但两者均未说明选定阈值下的误报率，且重复缺陷位置与掩码形状会人为提高重叠率。本文结合无分布上界阈值与配对-交叉空间检验，比较检测器得分位置与匹配缺陷掩码及其他图像掩码的差异，以差值衡量相对于经验随机重叠率的空间证据提升。在三个模态的120张点缺陷图像上评估三种检测器，共378次报警中230次与匹配掩码重叠。九个检测器-模态组合中，八组配对与交叉率相近；仅DINOv2-ASM在95%置信区间下具正向提升（0.259，0.159–0.347）。在独立的Magnetic Tile Defect数据集上，WRN50和ViT-B/16的提升分别为0.203（0.169–0.236）和0.231（0.202–0.262），单侧置换检验p=10⁻⁵。当交叉掩码限制为同类别时，提升仍为0.185和0.210。精确样本规划表明：150个校准正常样本仅支持1.98%及以上目标误报率的95%置信声明；若目标为1%，至少需299个正常样本。结果建议同时报告工作点性能与机会校正后的空间证据。
原文摘要 · Abstract (English)
Studies of industrial visual inspection commonly report the area under the receiver operating characteristic curve (AUROC) and the overlap between anomaly maps and defect masks. Neither measure specifies the false-alarm rate at a selected threshold, while recurrent defect locations and mask geometry can inflate overlap. We combine a distribution-free upper tolerance threshold with a paired-minus-crossed spatial test. This test compares each detector's score-contributing locations with the matched defect mask and with masks from other images; the difference in rates defines spatial-evidence lift relative to the empirical chance-overlap rate. We evaluate three detectors on 120 point-defect images from three ISP-AD modalities and three fixed data splits. Of 378 alarms, 230 overlap the matched mask. Paired and crossed rates are nevertheless similar in eight of nine detector--modality cells; only DINOv2--ASM has a positive 95\% bootstrap lower bound (lift 0.259, 95\% interval 0.159--0.347). On the independent Magnetic Tile Defect dataset, the same analysis gives lifts of 0.203 (0.169--0.236) for Wide ResNet-50 (WRN50) patch memory and 0.231 (0.202--0.262) for Vision Transformer B/16 (ViT-B/16) patch memory, with one-sided permutation $p=10^{-5}$ for both. When crossed masks are restricted to the same defect class, the lifts remain 0.185 and 0.210. Exact sample planning shows that, with 150 calibration normals, a 95\%-confidence distribution-free claim is supported only for target false-positive rates of 1.98\% or higher; a 1\% target requires at least 299 normals. The results support reporting operating-point performance and chance-corrected spatial evidence alongside AUROC and raw mask overlap.
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。