I am working with this face detection and recognition application and I don't have any past experience making same project. I don't know why it always show unknown even I already have 10 sample images of myself. Please someone a good understanding about face detection and EMGU can help me. This is my code:
using System;
using System.Collections.Generic;
using System.ComponentModel;
using System.Data;
using System.Drawing;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using System.Windows.Forms;
using Emgu.CV;
using Emgu.CV.Structure;
using Emgu.CV.Face;
using Emgu.CV.CvEnum;
using System.IO;
using System.Threading;
using System.Diagnostics;
namespace Version2
{
public partial class Form1 : Form
{
#region Variables
// for camera capture
private Capture videoCapture = null;
Mat frame = new Mat(); // use to store image data
private Image<Bgr, Byte> currentFrame = null; // current frame
private Image<Bgr, Byte> detectedFace = null; // detedted face
// cascade file for face detection
CascadeClassifier faceCasacdeClassifier = new CascadeClassifier(@"haarcascade_frontalface_default.xml");
// for face recognition
EigenFaceRecognizer recognizer;
List<Image<Gray, Byte>> TrainedFaces = new List<Image<Gray, byte>>();
List<int> PersonsLabes = new List<int>();
List<string> PersonsNames = new List<string>();
#endregion
public Form1()
{
InitializeComponent();
loadSavedFaceFiles();
}
private void loadSavedFaceFiles()
{
int ImagesCount = 0;
TrainedFaces.Clear();
PersonsLabes.Clear();
PersonsNames.Clear();
try
{
string path = Directory.GetCurrentDirectory() + @"\SavedFaces";
string[] files = Directory.GetFiles(path, "*.jpg", SearchOption.AllDirectories);
foreach (var file in files)
{
Image<Gray, byte> trainedImage = new Image<Gray, byte>(file).Resize(200, 200, Inter.Cubic);
CvInvoke.EqualizeHist(trainedImage, trainedImage);
TrainedFaces.Add(trainedImage);
PersonsLabes.Add(ImagesCount);
PersonsNames.Add(file.Split('\\').Last().Split('_')[0]);
ImagesCount++;
Debug.WriteLine(ImagesCount + ". " + file);
}
if (TrainedFaces.Count() > 0)
{
recognizer = new EigenFaceRecognizer(ImagesCount, double.PositiveInfinity);
recognizer.Train(TrainedFaces.ToArray(), PersonsLabes.ToArray());
}
}
catch (Exception ex)
{
MessageBox.Show("Error in Train Images: " + ex.Message);
}
}
private void processFrame(object sender, EventArgs e)
{
if (videoCapture != null && videoCapture.Ptr != IntPtr.Zero)
{
// takes a frame on the camera and converts it into an image object adjusting its size using picCamera size
videoCapture.Retrieve(frame, 0);
currentFrame = frame.ToImage<Bgr, Byte>().Resize(picCamera.Width, picCamera.Height, Inter.Cubic);
// display frame to the picture boxes
picCamera.Image = currentFrame.Bitmap;
picFaceDetection.Image = currentFrame.Bitmap;
// Convert from Bgr to Gray Image
Mat grayImage = new Mat();
CvInvoke.CvtColor(currentFrame, grayImage, ColorConversion.Bgr2Gray);
// Enhance the image to get better result
CvInvoke.EqualizeHist(grayImage, grayImage);
// detect faces
Rectangle[] faces = faceCasacdeClassifier.DetectMultiScale(grayImage, 1.1, 3, Size.Empty, Size.Empty);
labelFacesCount.Text = "Faces: " + faces.Length;
if (faces.Length > 0)
{
foreach (var face in faces)
{
// draws a rectagle on every face detected
Image<Bgr, Byte> boxedFrame = currentFrame.Clone();
CvInvoke.Rectangle(boxedFrame, face, new Bgr(Color.Red).MCvScalar, 2);
picFaceDetection.Image = boxedFrame.Bitmap;
// display cropped area of an image obect into pictureBox1
detectedFace = currentFrame.Convert<Bgr, Byte>();
detectedFace.ROI = face;
pictureBox1.SizeMode = PictureBoxSizeMode.StretchImage;
pictureBox1.Image = detectedFace.Bitmap;
// recognize face
if (PersonsNames.Count > 0)
{
Image<Gray, Byte> grayFaceResult = detectedFace.Convert<Gray, Byte>().Resize(200, 200, Inter.Cubic);
CvInvoke.EqualizeHist(grayFaceResult, grayFaceResult);
var result = recognizer.Predict(grayFaceResult);
pictureBox2.Image = grayFaceResult.Resize(pictureBox2.Width, pictureBox2.Height, Inter.Cubic).Bitmap;
pictureBox3.Image = TrainedFaces[result.Label].Resize(pictureBox3.Width, pictureBox3.Height, Inter.Cubic).Bitmap;
label1.Text = result.Label + ". " + result.Distance;
//Here results found known faces
if (result.Label != -1 && result.Distance < 2000)
{
CvInvoke.PutText(currentFrame, PersonsNames[result.Label], new Point(face.X - 2, face.Y - 2),
FontFace.HersheyComplex, 1.0, new Bgr(Color.Orange).MCvScalar);
CvInvoke.Rectangle(currentFrame, face, new Bgr(Color.Green).MCvScalar, 2);
}
//here results did not found any know faces
else
{
CvInvoke.PutText(currentFrame, "Unknown", new Point(face.X - 2, face.Y - 2),
FontFace.HersheyComplex, 1.0, new Bgr(Color.Orange).MCvScalar);
CvInvoke.Rectangle(currentFrame, face, new Bgr(Color.Red).MCvScalar, 2);
}
}
}
}
}
}
private void saveFace(object sender, EventArgs e)
{
string path = Directory.GetCurrentDirectory() + @"\SavedFaces";
if (!Directory.Exists(path))
{
Directory.CreateDirectory(path);
}
Task.Factory.StartNew(() => {
for (int i = 0; i < 10; i++)
{
//resize the image then saving it
detectedFace.Resize(200, 200, Inter.Cubic).Save(path + @"\" + txtPersonName.Text + "_" + DateTime.Now.ToString("dd-mm-yyyy-hh-mm-ss") + ".jpg");
Thread.Sleep(1000);
if (i + 1 == 10)
{
loadSavedFaceFiles();
Thread.Sleep(5000);
}
}
});
}
private void startCameraCapture(object sender, EventArgs e)
{
if (videoCapture != null) videoCapture.Dispose();
videoCapture = new Capture();
//videoCapture.ImageGrabbed += processFrame;
Application.Idle += processFrame;
// videoCapture.Start();
}
}
}
Thank you for those who can give a solution on this.
