通过遮蔽反向知识蒸馏提升异常检测精度
A Masked Reverse Knowledge Distillation Method Incorporating Global and Local Information for Image Anomaly Detection
- 用图像和特征级遮蔽，将重建转为修复任务
- 在MVTec数据集上达98.9%图像级、98.4%像素级AU-ROC
- 适合需要高精度异常定位的工业质检场景
知识蒸馏是有效的图像异常检测与定位方法，但其易过度泛化，主要源于输入与监督信号之间的相似性。为此，本文提出一种新方法——掩码反向知识蒸馏（MRKD）。通过图像级遮蔽（ILM）和特征级遮蔽（FLM），MRKD将图像重建任务转化为图像修复任务。其中，ILM通过区分输入与监督信号来捕捉全局信息；FLM引入合成特征级异常，确保学习表征包含足够局部信息。该方法增强了上下文建模能力，减少过度泛化。在广泛使用的MVTec异常检测数据集上的实验表明，MRKD取得优异性能：图像级98.9% AU-ROC，像素级98.4% AU-ROC，95.3% AU-PRO。大量消融实验验证了其在缓解过度泛化问题上的优势。
原文摘要 · Abstract (English)
Knowledge distillation is an effective image anomaly detection and localization scheme. However, a major drawback of this scheme is its tendency to overly generalize, primarily due to the similarities between input and supervisory signals. In order to address this issue, this paper introduces a novel technique called masked reverse knowledge distillation (MRKD). By employing image-level masking (ILM) and feature-level masking (FLM), MRKD transforms the task of image reconstruction into image restoration. Specifically, ILM helps to capture global information by differentiating input signals from supervisory signals. On the other hand, FLM incorporates synthetic feature-level anomalies to ensure that the learned representations contain sufficient local information. With these two strategies, MRKD is endowed with stronger image context capture capacity and is less likely to be overgeneralized. Experiments on the widely-used MVTec anomaly detection dataset demonstrate that MRKD achieves impressive performance: image-level 98.9% AU-ROC, pixel-level 98.4% AU-ROC, and 95.3% AU-PRO. In addition, extensive ablation experiments have validated the superiority of MRKD in mitigating the overgeneralization problem.
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。