测试C#调用OpenCvSharp和ViewFaceCore从摄像头中识别人脸

  学习了基于OpenCvSharp获取摄像头数据，同时学习了基于ViewFaceCore的人脸识别用法，将这两者结合即是从摄像头中识别人脸。本文测试测试C#调用OpenCvSharp和ViewFaceCore从摄像头中识别人脸，并进行人脸红框标记。
  新建Winform项目，在Nuget包管理器其中添加以下包：

Hompus.VideoInputDevices
DirectShowLib.Standard
OpenCvSharp4
OpenCvSharp4.Extensions
OpenCvSharp4.runtime.win
ViewFaceCore
ViewFaceCore.all_models
ViewFaceCore.Extension.SkiaSharp
ViewFaceCore.runtime.win.x64

  关键代码如下所示，通过OpenCvSharp从摄像头获取图片后，调用ViewFaceCore模块识别图片中的人脸，并用红框标记人脸，最后显示图片。

m_capture.Read(m_image);

if (m_image.Empty())
    return;

Bitmap image= OpenCvSharp.Extensions.BitmapConverter.ToBitmap(m_image);

var bitmap = SKBitmap.Decode(m_image.ToBytes());
FaceDetector faceDetector = new FaceDetector();

FaceLandmarkConfig config = new FaceLandmarkConfig();
config.MarkType = MarkType.Normal;

FaceLandmarker faceMark = new FaceLandmarker(config);
           
FaceInfo[] infos = faceDetector.Detect(bitmap);

using(Graphics graphics = Graphics.FromImage(image))
{
    for (int i = 0; i < infos.Length; i++)
    {
        graphics.DrawRectangle(Pens.Red,
                infos[i].Location.X,
                infos[i].Location.Y,
                infos[i].Location.Width,
                infos[i].Location.Height);
    }
} 

faceDetector.Dispose();
faceMark.Dispose();

pictureBox1.Image = image;

  最后是程序运行效果，在手机的B站上找了“花都兄弟”的视频，对着摄像头播放，可以看到能够识别到人脸并标记。

参考文献：
[1]https://github.com/ViewFaceCore/ViewFaceCore
[2]https://github.com/shimat/opencvsharp
[3]https://www.bilibili.com/video/BV1Je411p7BQ/?spm_id_from=333.337.search-card.all.click