Skip to content

mathworks/visual-Inspection-examples-using-PatchCore-and-FastFlow-

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 

Repository files navigation

PatchCoreとFastFlowを使った外観検査AIモデル開発のサンプル

About this sample

このサンプルは2023年3月30日に実施したWebinar「いまからでも遅くない外観検査〜ワークフロー概説と最新アルゴリズム適用例」の後半のセッション「外観検査を効率化する最新アルゴリズム “PatchCore”の原理と適用例」 で紹介したPatchCoreのサンプル、そしてFastFlowのサンプルです。 元のサンプルである「Detect Image Anomalies Using Explainable FCDD Network 」から最低限の変更に留めて作成していますので、解説のテキストとコードが一致していない箇所がある点ご留意ください。 変更点はWebinar中に明示していますので参考にしてください。

How to run samples

PatchCoreのサンプル
MATLAB同梱のPillデータセットを使って、PatchCoreによる外観検査AIモデルを学習、検証することができます。
Demosフォルダにある"PatchCore_using_Pill_data_Example.mlx"を開いて、実行してください。

FastFlowのサンプル
お手持ちのデータを使ってFastFlowによる外観検査AIモデルを学習、検証することができます。
Demosフォルダにある"FastFlow_using_mydata_Example.mlx"を開いて、dataDirやimageDirを適宜変更し実行してください。
変更箇所はWebinarを参考にしてください。

Recording

MathWorks® Webinar いまからでも遅くない外観検査〜ワークフロー概説と最新アルゴリズム適用例アーカイブページ

Required Products

Products:

  • MATLAB™
  • Image Processing Toolbox™
  • Computer Vision Toolbox™
  • Deep Learning Toolbox™
    (※いずれもR2023a以降)

Add-on:

  • Computer Vision Toolbox Automated Visual Inspection Library(Computer Vision Toolboxのライセンスがあれば無償ダウンロード可能です)
  • Deep Learning Toolbox Model for ResNet-18 Network

References

PatchCore: proposed in Roth et al. (2021), Towards Total Recall in Industrial Anomaly Detection
FastFlow: proposed in Yu et al. (2021), FastFlow: Unsupervised Anomaly Detection and Localization via 2D Normalizing Flows

Releases

No releases published

Packages

No packages published