发表机构
KAIST; Hanyang University; Samsung Display（韩国科学技术院; 汉阳大学; 三星显示）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对多光照/多聚焦显示图像异常检测，提出基于对比蒸馏（MCD）和混合模块的方法，在MMdAD数据集上显著超越现有方法。
AI 中文摘要
本文研究了多光照和多聚焦显示图像中的自动异常检测问题。显示表面上的细微缺陷在RGB图像中难以被发现，且仅用正常数据训练的模型也难以识别这些缺陷。为解决这一问题，我们提出了一种新颖的对比学习方案，用于基于知识蒸馏的异常检测。在我们的框架中，采用多分辨率知识蒸馏（MKD）作为基线，该方法通过测量教师网络与学生网络之间的特征相似性来运行。基于MKD，我们提出了一种新颖的对比学习方法，即多分辨率对比蒸馏（MCD），该方法不需要带有锚点的正/负样本对，而是通过拉近或推远教师与学生特征之间的距离来运行。此外，我们提出了一个混合模块，将多通道信息转换并聚合到MCD的三通道输入层中。在我们收集的多光照和多聚焦显示图像异常检测数据集（MMdAD）上，我们提出的方法在AUROC和准确率指标上均显著优于现有的最先进方法。
英文摘要
In this paper, we tackle automatic anomaly detection in multi-illumination and multi-focus display images. The minute defects on the display surface are hard to spot out in RGB images and by a model trained with only normal data. To address this, we propose a novel contrastive learning scheme for knowledge distillation-based anomaly detection. In our framework, Multiresolution Knowledge Distillation (MKD) is adopted as a baseline, which operates by measuring feature similarities between the teacher and student networks. Based on MKD, we propose a novel contrastive learning method, namely Multiresolution Contrastive Distillation (MCD), which does not require positive/negative pairs with an anchor but operates by pulling/pushing the distance between the teacher and student features. Furthermore, we propose the blending module that transforms and aggregate multi-channel information to the three-channel input layer of MCD. Our proposed method significantly outperforms competitive state-of-the-art methods in both AUROC and accuracy metrics on the collected Multi-illumination and Multi-focus display image dataset for Anomaly Detection (MMdAD).
CommentsMVA 2023 (Oral)
Journal ref2023 18th International Conference on Machine Vision and Applications (MVA)
DOI：10.23919/MVA57639.2023.10215808
