Blog post: Real-time face recognition with Microsoft Cognitive Services

YouTube video: .NET User Group: Real-time Face Recognition With Microsoft Cognitive Services

Cognitive Studio (in progress): https://cognitivestudio.dev/face/intro

RealTimeFaceApi

This is a demo project showing how to use Face API in Cognitive Services with OpenCV.

The demo has several parts to allow real-time facial recognition:

• Get frames from web cam/video with help of OpenCV
• Offline face-detection with help of OpenCV
• Filter out faces that are too small
• Determine if scene has changed and we should recognize new faces
• Sending and identifying faces with help of Microsoft Cognitive Services

Setup

In RealTimeFaceApi/Program.cs configure FaceSubscriptionKey and FaceGroupId from Microsoft Cognitive Services. You'll need to upload and train data in MS Face API.

Run with web came

Make sure there are no command line arguments.

If you have multiple web cams, you can change the camera with cameraIndex.

// Otherwise use the webcam.
capture = InitializeCapture(/* camera index */ 1);

You can run app from console:

cd RealTimeFaceApi.Cmd
dotnet run

Run with a video file

Add path to file. In Visual Studio, go to RealTimeFaceApi.Cmd properties, Debug and under Application arguments: add the path to the video.

You can also run from console:

cd RealTimeFaceApi.Cmd
dotnet run -- "C:\Users\JK\Downloads\Real-time Face Recognition With Microsoft Cognitive Services.mp4"