转自AI Studio,原文链接: 面向全召回率的工业异常检测 - 飞桨AI StudioPatchCore: Towards Total Recall in Industrial Anomaly Detection 1. 简介 本项目基于PaddlePaddle框架复现了PatchCore算法,并在MvTec数…
PatchCore 论文: PatchCore: Towards Total Recall in Industrial Anomaly Detection 参考repo: anomalib 在此非常感谢openvineo贡献的anomalib项目,提高了本repo复现论文的效率。 本项目为第六届论文复现赛异常检测项目。 感谢百度提供的算力支持。 2 复现精度 按复现考核标准,使用resnet18 10%特征数 在MVTec AD...
First of all, install all the required frameworks and import our repository. Then, you need tochoose which backbone you want to useasPretrained Encoderto run the PatchCore method. If you want to run the vanilla version of PatchCore, useWideResNet50as backbone. If you want to test the CLIP...