Face Detection in a Video Frames

In this tutorial, you'll learn how to detect faces in a video frames from your camera using Capturer object from Face SDK API. Detected faces are highlighted with a green rectangle.

Besides Face SDK and Visual Studio 2019, you'll need a camera connected to your PC (for example, a webcam). You can build and run this project on Windows.

You can find the tutorial project in Face SDK: examples/tutorials/csharp/FirstApp

Prepare project

1. Run Visual Studio and create new project New > Project > Windows Forms application (.NET FrameWork) > Next. Then you can to mention the name and location of our project. You will see a Form Designer displayed in Visual Studio. In this Form Designer, you will start building a Windows Forms application.

2. Double-click the left mouse button on the form. This will create a new Form1_Load method in the Form1.cs. It will be needed further.

3. Open Build > Configuration Manager. In the field Platform in the combobox select <New...> and choose platform x64.

   

   Click OK > Close.

4. Open the Form Designer and add the PictureBox component to the form (located in View > Toolbox > All Windows Forms). For the created pictureBox1 click Dock in parent container and select Size Mode Zoom.

   

Getting a video from a webcam

1. To work with the webcam, we will use the Emgu.CV and Emgu.CV.Bitmap libraries. To install the package use NuGet Package Manager (Tools > NuGet Package Manager > Manage Nuget Packages for Solution). Select the Browse tab, find the packages and install them.

   

2. Open the Form1.cs code editor and add imports to the beginning of the code:

   using Emgu.CV;
   using Emgu.CV.Structure;

3. Add the GetFrame method to the class, modify Form1_Load. Also adds the variables camera and camera_id with device number:

   namespace FirstApp
   {
       public partial class Form1 : Form
       {
           private int camera_id = 0;
           private VideoCapture camera;
           ...
           private void GetFrame(object sender, EventArgs e)
           {
               Image<Bgr, byte> image = camera.QueryFrame().ToImage<Bgr, byte>();
               pictureBox1.Image = image.ToBitmap();
           }
   
           private void Form1_Load(object sender, EventArgs e)
           {
               camera = new VideoCapture(camera_id);
               if (camera.IsOpened)
               {
                   Application.Idle += GetFrame;
               }else
               {
                   MessageBox.Show("Camera not opened!", "Camera error", MessageBoxButtons.OK, MessageBoxIcon.Error);
                   Application.Exit();
               }
           }
           ...
       }
   }

4. Now you can build and start program Build > Run Code Analysis on Solution and make sure that the video from the camera is displayed.

   

Adding FaceSDK to a project

1. Open View > Solution Explorer > References (right-click) > Add Reference -> Browse and add FacerecCSharpWrapper.dll from <facesdk_root_dir>\bin folder. Click OK.

2. To automatically copy libraries into your project open Project > {PROJECT_NAME} Properties... > Build Events > Edit Pre-build... and add lines:

   set facesdk_root_dir=<facesdk_root_dir>
   xcopy /Y "%facesdk_bin%\bin\facerec.dll" "$(TargetDir)"
   xcopy /Y "%facesdk_bin%\bin\sense4.dll" "$(TargetDir)"
   xcopy /Y "%facesdk_bin%\bin\tensorflow.dll" "$(TargetDir)"

   Note: replace <facesdk_root_dir> with the path where the installed FaceSDK is located.

   

   Click OK.

3. In Form1.cs add imports:

   using System.Runtime.InteropServices;
   using VDT.FaceRecognition.SDK;

4. Add private variables into Form1 class:

   private String faceSDKRootDir = "<facesdk_root_dir>";
   private Capturer capturer;

   Note: replace <facesdk_root_dir> with the path where the installed FaceSDK is located.

5. Before the line Application.Idle ... add the creation of the FaceSDK service and the capturer to the Form1_Load method:

   FacerecService service = FacerecService.createService(faceSDKRootDir + "\\conf\\facerec", "");
   FacerecService.Config capturerConfig = new FacerecService.Config("fda_tracker_capturer_uld.xml");
   capturer = service.createCapturer(capturerConfig);

6. The final step is to add a method that will detect the face in the image and draw a frame around the face.

   private void drawDetections(Image<Bgr, byte> image)
   {
       Mat frame_m = image.Mat.Clone();
       byte[] data = new byte[frame_m.Total.ToInt32() * frame_m.NumberOfChannels];
       Marshal.Copy(frame_m.DataPointer, data, 0, (int)data.Length);
       RawImage ri_frame = new RawImage(frame_m.Width, frame_m.Height, RawImage.Format.FORMAT_BGR, data);
       List<RawSample> detected = capturer.capture(ri_frame);
       foreach (RawSample sample in detected)
       {
           RawSample.Rectangle rect = sample.getRectangle();
           image.Draw(new Rectangle((int)rect.x,
                                    (int)rect.y,
                                    (int)rect.width,
                                    (int)rect.height),
                      new Bgr(0, 255, 0),
                      2);
       }
   }

   Also add its call to the GetFrame method:

   private void GetFrame(object sender, EventArgs e)
   {
       Image<Bgr, byte> image = camera.QueryFrame().ToImage<Bgr, byte>();
       drawDetections(image);
       pictureBox1.Image = image.ToBitmap();
   }

7. Now you can launch the application and check how it works.