README.md

---
title: MVTec Anomaly Detection Demo
emoji: 🔍
colorFrom: blue   
colorTo: blue
sdk: gradio
sdk_version: "4.44.0"
app_file: app.py
pinned: false
license: mit
---

# MVTec Anomaly Detection Demo

Upload an image or pick a sample to detect anomalies using PaDiM / PatchCore / FastFlow models trained on the MVTec AD benchmark.

## Models
- **PaDiM** — Patch Distribution Modeling (multivariate Gaussian per patch)
- **PatchCore** — Memory bank + K-Nearest Neighbors
- **FastFlow** — Normalising flows on CNN features

## Dataset
[MVTec Anomaly Detection Dataset](https://www.mvtec.com/company/research/datasets/mvtec-ad)