Skip to content
Projects

georg-jung/explain-face-rec

View on GitHub

georg-jung/explain-face-rec has 33 GitHub stars and 8 forks. GitHot measured 0 stars in 24 hours. Last checked 2026-11-16 03:10 UTC. Last push 2026-11-13.

How these counts are measured

What you can do

  • Interactively learn how facial recognition works.
  • Use your own images OR get started right away with sample images.
  • Apply advanced face AI on your own machine.
  • Follow along with hands-on code examples.

Measured star growth

in 24h
0
in 7 days
Update delayed
in 15 days
Update delayed
in 1 month
Update delayed

Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.

About

Understand how modern technology works with faces, from scratch. You don't need to be an AI expert. If you know how to program, you will be able to apply state-of-the-art methods yourself, locally on your device and for free. This is an interactive tutorial – you can bring your own pictures if you like to.

AI

README.md

Original

logo

Interactive Face Recognition Tutorial

facerec.gjung.com Google Play Microsoft Store

Understand how modern technology works with faces, from scratch. You don't need to be an AI expert. If you know how to program, you will be able to apply state-of-the-art methods yourself, locally on your device and for free. This is an interactive tutorial – you can bring your own pictures if you like to.

Features

  • Interactively learn how facial recognition works.
  • Use your own images OR get started right away with sample images.
  • Apply advanced face AI on your own machine.
  • Follow along with hands-on code examples.
  • Understand how to use AI with .NET and Blazor.
  • Experiment to see how parameters change results.
  • Apply what you learn to your own production applications.
  • Free, local, no dependency on cloud services.
  • Uses the MIT-licensed FaceAiSharp library.

Get started

Get started using the web app facerec.gjung.com or by downloading this tutorial as an app:

Get it on Google Play Get it from Microsoft

Google Play and the Google Play logo are trademarks of Google LLC.

Using Docker

docker run -p 8080:8080 ghcr.io/georg-jung/explain-face-rec:latest

As soon as your container is running, you can access the tutorial at localhost:8080.

Building the MAUI apps

The native apps use stable .NET 10 and MAUI 10. Android targets Android 16 (API 36) and supports Android 7.0 (API 24) or newer; Windows targets the Windows 10 SDK 19041. Install a stable .NET 10 SDK compatible with global.json, and use Git LFS to retrieve the sample images (git lfs pull). The AI models are restored through NuGet.

For Android, install Microsoft OpenJDK 21 and set JAVA_HOME and ANDROID_HOME to your JDK and Android SDK directories. Then run from the repository root:

dotnet workload install maui-android
dotnet build src/BlazorFace.Maui/BlazorFace.Maui.csproj -t:InstallAndroidDependencies -f net10.0-android36.0 -p:MauiTargetFrameworks=net10.0-android36.0 -p:AcceptAndroidSdkLicenses=true
dotnet build src/BlazorFace.Maui/BlazorFace.Maui.csproj -f net10.0-android36.0 -p:MauiTargetFrameworks=net10.0-android36.0 -c Release

InstallAndroidDependencies installs the SDK components required by the project's target API. See Microsoft's Android dependency setup for initial SDK installation and custom paths.

On Windows, install the Windows development tools and SDK with Visual Studio's .NET MAUI workload, then build just the Windows target:

dotnet workload install maui-windows
dotnet build src/BlazorFace.Maui/BlazorFace.Maui.csproj -f net10.0-windows10.0.19041.0 -p:MauiTargetFrameworks=net10.0-windows10.0.19041.0 -p:NoAndroid=true -c Release

MauiTargetFrameworks selects the app platform without overriding the frameworks of its referenced projects. NoAndroid=true also skips the shared library's Android target when building on Windows.

The MAUI workflow validates pull requests and manual runs without production signing credentials. Published GitHub releases additionally produce a signed Android App Bundle using the existing KEYSTORE_FILE_BASE64 and KEYSTORE_PASSWORD secrets and key alias key. Passwords use temporary files with the file: syntax supported by AAB signing. Windows builds produce unsigned packages for Store submission. Artifacts are attached to the workflow run; the workflow does not submit apps to either store.

Nerdbank.GitVersioning derives Android's version code from version.json and Git history.

Credits

Created in the context of Georg Jung's master thesis, supervised by Dr. Guido Rößling. Thanks for the supervision and the great support!