CAMSHIFT in EmguCV

Hey Assalam-O-Alikum..

Atlast im successsful in making a program that can track object using CAMSHIFT algorithim with the help of EmguCV.. EmguCv is basically an openCV wrapper class.. it can be downloaded from EmguCv website.. ( ) it is free and open source... :)

here is a piece of code how i accomplished my MISSION.. :P :)




namespace WindowsFormsApplication1
{
public partial class frmCamShift : Form
{

private Capture cap;
private HaarCascade haar;

private Image faceS;

private static RangeF range = new RangeF(0, 180);
private DenseHistogram hist = new DenseHistogram(16, range);

private Image hue = null;
private Image mask = null;
private Image backproject = null;
private Image hsv = null;

private IntPtr[] img = null;
private Rectangle trackwin;
private MCvConnectedComp trackcomp = new MCvConnectedComp();
private MCvBox2D trackbox = new MCvBox2D();

private bool isTrack = false;

public frmCamShift()
{
InitializeComponent();
}

private void frmCamShift_Load(object sender, EventArgs e)
{
cap = new Capture(0);
//haar = new HaarCascade("haarcascade_frontalface_alt.xml");
haar = new HaarCascade("haarcascade.xml");
cap.FlipHorizontal = true;
}

private void timer1_Tick(object sender, EventArgs e)
{
using(Image nextframe=cap.QueryFrame())
{
if (nextframe != null)
{
if (isTrack == false)
{
Image grayframe = nextframe.Convert();
grayframe._EqualizeHist();

var faces = grayframe.DetectHaarCascade(haar, 1.4, 4, HAAR_DETECTION_TYPE.FIND_BIGGEST_OBJECT | HAAR_DETECTION_TYPE.DO_CANNY_PRUNING, new Size(40, 40))[0];

hsv = new Image(grayframe.Width, grayframe.Height);
hsv = nextframe.Convert();
hsv._EqualizeHist();

hue = new Image(grayframe.Width, grayframe.Height);
mask = new Image(grayframe.Width, grayframe.Height);
backproject = new Image(grayframe.Width, grayframe.Height);

Emgu.CV.CvInvoke.cvInRangeS(hsv, new MCvScalar(0, 30, Math.Min(10, 255), 0), new MCvScalar(180, 256, Math.Max(10, 255), 0), mask);
Emgu.CV.CvInvoke.cvSplit(hsv, hue, IntPtr.Zero, IntPtr.Zero, IntPtr.Zero);

picHue.Image = hue.ToBitmap();


foreach (var face in faces)
{

// Rectangle roi = new Rectangle(face.rect.X + face.rect.Width / 4, face.rect.Y + face.rect.Height / 4, face.rect.Width / 2, face.rect.Height / 2);
// Rectangle roi = new Rectangle(face.rect.X, face.rect.Y, face.rect.Width / 2, face.rect.Height / 2);

Emgu.CV.CvInvoke.cvSetImageROI(hue, face.rect);
Emgu.CV.CvInvoke.cvSetImageROI(mask, face.rect);

nextframe.Draw(face.rect, new Bgr(0, double.MaxValue, 1), 2);
picMask.Image = mask.ToBitmap();
trackwin = face.rect;

}
img = new IntPtr[1]
{
hue
};

Emgu.CV.CvInvoke.cvCalcHist(img, hist, false, mask);

Emgu.CV.CvInvoke.cvResetImageROI(hue);
Emgu.CV.CvInvoke.cvResetImageROI(mask);

CapImg.Image = nextframe.ToBitmap();
isTrack = true;
// isTrack = true;
}
else
{
if (trackwin != null)
{
hsv = nextframe.Convert();
Emgu.CV.CvInvoke.cvInRangeS(hsv, new MCvScalar(0, 30, 10, 0), new MCvScalar(180, 256, 256, 0), mask);
Emgu.CV.CvInvoke.cvSplit(hsv, hue, IntPtr.Zero, IntPtr.Zero, IntPtr.Zero);
picMask.Image = mask.ToBitmap();
picHue.Image = hue.ToBitmap();

}

img = new IntPtr[1]
{
hue
};

Emgu.CV.CvInvoke.cvCalcBackProject(img, backproject, hist);
Emgu.CV.CvInvoke.cvAnd(backproject, mask, backproject, IntPtr.Zero);

Image grayframe = nextframe.Convert();
grayframe._EqualizeHist();


var faces = grayframe.DetectHaarCascade(haar, 1.4, 4, HAAR_DETECTION_TYPE.FIND_BIGGEST_OBJECT | HAAR_DETECTION_TYPE.DO_CANNY_PRUNING, new Size(40, 40))[0];
foreach (var face in faces)
{
nextframe.Draw(face.rect, new Bgr(Color.Black), 2);
}


if (trackwin.Width == 0) trackwin.Width = 40;
if (trackwin.Height == 0) trackwin.Height = 40;

Emgu.CV.CvInvoke.cvCamShift(backproject, trackwin, new MCvTermCriteria(10,0.1), out trackcomp, out trackbox);
trackwin = trackcomp.rect;

// CvInvoke.cvEllipseBox(nextframe, trackbox, new MCvScalar(0, 255, 0), 2, LINE_TYPE.CV_AA, 0);


nextframe.Draw(trackwin, new Bgr(Color.Blue), 3);
CapImg.Image = nextframe.ToBitmap();
faceS = nextframe.Copy(trackwin);
picFace.Image = faceS.ToBitmap();

}
}


}
}

private void cmdRefresh_Click(object sender, EventArgs e)
{
//timer1.Enabled = false;
//isTrack = false;
//trackwin.Width = 0;
//trackwin.Height = 0;
//hue.Dispose();
//mask.Dispose();
//hsv.Dispose();
//backproject.Dispose();
//img = null;
//timer1.Enabled = true;
}
}
}


i wish i could put some comments on it but unfortunatly i cant do that right now..

here is a video of how it worked.. my Webcam gives really really BAD result :(.. just 2-4 fps... maybe its because of my slow processor and freaking webcam.. any way, it was big achievement for me as i was working on it for a month... :)



my hair style looks Ugly.. LOlZzzzXXX

it detects the face using haar classifiers and then tracks it using CAMShift algorithim.. im trying to do some more work on it as i want to disply in my univesity's IT fair, so NO cheating allowed till MAY 2010.. :)

You can find another post of mine here face and facial parts detection using emgucv and haar classifiers

Dont forget to comment.. :)

Ca ya Later..
Jazak-Allah