public bool TrainRecognizer()
{
var allFaces = _dataStoreAccess.CallFaces("ALL_USERS");
if (allFaces != null)
{
var faceImages = new Image <Gray, byte> [allFaces.Count];
var faceLabels = new int[allFaces.Count];
for (int i = 0; i < allFaces.Count; i++)
{
Stream stream = new MemoryStream();
stream.Write(allFaces[i].Image, 0, allFaces[i].Image.Length);
var faceImage = new Image <Gray, byte>(new Bitmap(stream));
faceImages[i] = faceImage.Resize(200, 200, Inter.Cubic);
faceLabels[i] = allFaces[i].UserId;
}
//LBPHFaceRecognizer--------------------------
_LBPHFaceRecognizer.Train(faceImages, faceLabels);
_LBPHFaceRecognizer.Save(_recognizerFilePath);
////EigenFaceRecognizer-------------------------
//_faceRecognizer.Train(faceImages, faceLabels);
//_faceRecognizer.Save(_recognizerFilePath);
}
return(true);
}
// Trains a recognizer using the labeled images in inputDir and saves the result
// to outputPath
static void TrainRecognizer(string inputDir, string outputPath)
{
var imageFiles = Directory.EnumerateFiles(inputDir).ToList();
var images = new Image <Gray, byte> [imageFiles.Count];
var labels = new int[imageFiles.Count];
// Load each file and it's label
Console.WriteLine("Loading data...");
int i = 0;
foreach (var imageFile in imageFiles)
{
var label = Path.GetFileNameWithoutExtension(imageFile).Split('_')[0];
labels[i] = int.Parse(label);
images[i] = CvInvoke.Imread(imageFile, Emgu.CV.CvEnum.LoadImageType.Grayscale).ToImage <Gray, byte>();
i++;
}
Console.WriteLine("Loading complete.");
// Train the recognizer and save the result
Console.WriteLine("Training...");
LBPHFaceRecognizer recognizer = new LBPHFaceRecognizer();
recognizer.Train(images, labels);
Console.WriteLine("Training done. Saving results...");
recognizer.Save(outputPath);
}
private void Train()
{
var inputImages = new List <Image <Gray, byte> >();
var labelsList = new List <int>();
foreach (var photo in PhotoContext.Photos.Where(x => x.Id != 1).ToList())
{
MemoryStream str = new MemoryStream(photo.Face.Image);
Bitmap b = new Bitmap(Image.FromStream(str));
Image <Gray, byte> img = new Image <Gray, byte>(b);
img._EqualizeHist();
inputImages.Add(img);
labelsList.Add(photo.OwnerId);
}
DataPath = Server.MapPath("~/App_Data/Faces");
FaceRecognizer.Train(inputImages.ToArray(), labelsList.ToArray());
FaceRecognizer.Save(DataPath);
}
public void Save(string filename)
{
recognizer.Save(filename);
string path = Path.GetDirectoryName(filename);
FileStream labels = File.OpenWrite(Path.Combine(path, "Labels.xml"));
using (XmlWriter writer = XmlWriter.Create(labels))
{
writer.WriteStartDocument();
writer.WriteStartElement("labels");
for (int i = 0; i < names.Count; ++i)
{
writer.WriteStartElement("label");
writer.WriteElementString("name", names[i]);
writer.WriteEndElement();
}
writer.WriteEndElement();
writer.WriteEndDocument();
}
labels.Close();
}
/***
* Function: private void TrainMachine(FaceIdentity faceIdentity, String name)
* Parameter(s): FaceIdentity faceIdentity
* Privilege of the face that is being trained to store in the ASSET_INDEX.dat file.
* String name
* The name of the individual that is being trained; currently, it is not used, but it exists so that in the next version,
* the machine can be more customized.
* Return Value: void
***/
private void TrainMachine(FaceIdentity faceIdentity, String name)
{
// Notify the user that training has begun.
MessageBox.Show(trainingBeginning, this.Title, MessageBoxButton.OK);
String[] fileList = Directory.GetFiles(FileUtilities.TrainingDirectoryName);
List <Mat> matList = new List <Mat>();
foreach (String file in fileList)
{
matList.Add(new Mat(file, Emgu.CV.CvEnum.LoadImageType.Unchanged));
}
List <Image <Gray, Byte> > list = new List <Image <Gray, Byte> >();
// Detect each face in each image.
foreach (Mat mat in matList)
{
Image <Gray, Byte> image = mat.ToImage <Gray, Byte>().Resize(1 / (double)scale_factor, Emgu.CV.CvEnum.Inter.Cubic);
Rectangle[] faceList = faceClassifier.DetectMultiScale(image);
foreach (Rectangle rect in faceList)
{
list.Add(image.Copy(rect).Convert <Gray, Byte>());
}
}
// Make sure that there is at least one face to train.
if (list.Count() == 0)
{
PanicAndTerminateProgram();
}
// If a height exists in the CORE_IMAGE_DATA.dat file, resize to that, useful for future training.
int height = facialRecognitionHeight == 0 ? list[0].Height * scale_factor : facialRecognitionHeight;
int width = facialRecognitionWidth == 0 ? list[0].Width * scale_factor : facialRecognitionWidth;
if (facialRecognitionHeight == 0 || facialRecognitionWidth == 0)
{
List <String> lines = new List <String>();
lines.Add(height + "|" + width);
File.WriteAllLines(FileUtilities.DirectoryName + "\\" + FileUtilities.CoreImageData, lines.ToArray());
}
List <Image <Gray, Byte> > listFinal = new List <Image <Gray, Byte> >();
foreach (Image <Gray, Byte> image in list)
{
listFinal.Add(image.Resize(width, height, Emgu.CV.CvEnum.Inter.Cubic));
}
List <int> labelList = new List <int>();
int integer = 0;
String prefix = "";
String ident = "";
if (faceIdentity == FaceIdentity.FaceAdmin)
{
integer = 0;
prefix = FileUtilities.AdminTrainedPrefix;
ident = adminIdentifier;
}
else if (faceIdentity == FaceIdentity.FaceAsset)
{
integer = 2;
prefix = FileUtilities.AssetTrainedPrefix;
ident = auxAdminIdentifier;
}
else if (faceIdentity == FaceIdentity.FaceAuxAdmin)
{
integer = 1;
prefix = FileUtilities.AuxAdminTrainedPrefix;
ident = assetIdentifier;
}
else
{
PanicAndTerminateProgram();
}
for (int i = 0; i < list.Count(); i++)
{
labelList.Add(integer);
}
// Train the machine and write its trained state to a file.
LBPHFaceRecognizer lbphFaceRecognizer = new LBPHFaceRecognizer();
lbphFaceRecognizer.Train <Gray, Byte>(listFinal.ToArray(), labelList.ToArray());
Directory.Delete(FileUtilities.TrainingDirectoryName, true);
String temp = categories[integer];
String fname = FileUtilities.DirectoryName + "\\" + prefix + temp.ToUpper().Replace(' ', '_') + FileUtilities.FileExtension;
lbphFaceRecognizer.Save(fname);
// Write everything to the ASSET_INDEX.dat file.
FileUtilities.TrainingDirectoryDeletion();
List <String> aboutTraining = new List <String>();
aboutTraining.Add(name + "^" + ident + "^" + fname);
File.AppendAllLines(FileUtilities.DirectoryName + "\\" + FileUtilities.AssetIndexData, aboutTraining.ToArray());
// Notify the used that training has ended.
MessageBox.Show(trainingEnded, this.Title, MessageBoxButton.OK);
}
private void button1_Click(object sender, EventArgs e)
{
if (comboBoxAlgorithm.Text == "EigenFaces")
{
try
{
string dataDirectory = Directory.GetCurrentDirectory() + "\\TrainedFaces";
string[] files = Directory.GetFiles(dataDirectory, "*.jpg", SearchOption.AllDirectories);
eigenTrainedImageCounter = 0;
foreach (var file in files)
{
Image <Bgr, Byte> TrainedImage = new Image <Bgr, Byte>(file);
if (eqHisChecked.Checked == true)
{
TrainedImage._EqualizeHist();
}
eigenTrainingImages.Add(TrainedImage.Convert <Gray, Byte>());
eigenlabels.Add(fileName(file));
eigenIntlabels.Add(eigenTrainedImageCounter);
eigenTrainedImageCounter++;
richTextBox1.Text += fileName(file) + "\n";
}
/*
* //TermCriteria for face recognition with numbers of trained images like maxIteration
* MCvTermCriteria termCrit = new MCvTermCriteria(eigenTrainedImageCounter, 0.001);
*
* //Eigen face recognizer
* eigenObjRecognizer=new EigenObjectRecognizer(
* eigenTrainingImages.ToArray(),
* eigenlabels.ToArray(),
* 3000,
* ref termCrit);
*/
eigenFaceRecognizer = new EigenFaceRecognizer(eigenTrainedImageCounter, 2000);
eigenFaceRecognizer.Train(eigenTrainingImages.ToArray(), eigenIntlabels.ToArray());
//eigenFaceRecognizer.Save(dataDirectory + "\\trainedDataEigen.dat");
}
catch (Exception ex)
{
MessageBox.Show(ex.ToString());
MessageBox.Show("Nothing in binary database, please add at least a face(Simply train the prototype with the Add Face Button).", "Triained faces load", MessageBoxButtons.OK, MessageBoxIcon.Exclamation);
}
}
else if (comboBoxAlgorithm.Text == "FisherFaces")
{
try
{
string dataDirectory = Directory.GetCurrentDirectory() + "\\TrainedFaces";
string[] files = Directory.GetFiles(dataDirectory, "*.jpg", SearchOption.AllDirectories);
fisherTrainedImageCounter = 0;
foreach (var file in files)
{
Image <Bgr, Byte> TrainedImage = new Image <Bgr, Byte>(file);
fisherTrainingImages.Add(TrainedImage.Convert <Gray, Byte>());
if (eqHisChecked.Checked == true)
{
TrainedImage._EqualizeHist();
}
fisherlabels.Add(fileName(file));
fisherIntlabels.Add(fisherTrainedImageCounter);
fisherTrainedImageCounter++;
richTextBox1.Text += fileName(file) + "\n";
}
fisherFaceRecognizer = new FisherFaceRecognizer(fisherTrainedImageCounter, 2000);
fisherFaceRecognizer.Train(fisherTrainingImages.ToArray(), fisherIntlabels.ToArray());
//fisherFaceRecognizer.Save(dataDirectory + "\\trainedDataFisher.dat");
}
catch (Exception ex)
{
MessageBox.Show(ex.ToString());
MessageBox.Show("Nothing in binary database, please add at least a face(Simply train the prototype with the Add Face Button).", "Triained faces load", MessageBoxButtons.OK, MessageBoxIcon.Exclamation);
}
}
else if (comboBoxAlgorithm.Text == "LBPHFaces")
{
try
{
string dataDirectory = Directory.GetCurrentDirectory() + "\\TrainedFaces";
string[] files = Directory.GetFiles(dataDirectory, "*.jpg", SearchOption.AllDirectories);
lbphTrainedImageCounter = 0;
foreach (var file in files)
{
Image <Bgr, Byte> TrainedImage = new Image <Bgr, Byte>(file);
if (eqHisChecked.Checked == true)
{
TrainedImage._EqualizeHist();
}
lbphTrainingImages.Add(TrainedImage.Convert <Gray, Byte>());
lbphlabels.Add(fileName(file));
lbphIntlabels.Add(lbphTrainedImageCounter);
lbphTrainedImageCounter++;
richTextBox1.Text += fileName(file) + "\n";
}
lbphFaceRecognizer = new LBPHFaceRecognizer(1, 8, 8, 8, 400);
lbphFaceRecognizer.Train(lbphTrainingImages.ToArray(), lbphIntlabels.ToArray());
lbphFaceRecognizer.Save(dataDirectory + "\\trainedDataLBPH.dat");
}
catch (Exception ex)
{
MessageBox.Show(ex.ToString());
MessageBox.Show("Nothing in binary database, please add at least a face(Simply train the prototype with the Add Face Button).", "Triained faces load", MessageBoxButtons.OK, MessageBoxIcon.Exclamation);
}
}
}
