Emgu CV - Skeleton of an Image in EmguCV (Skeletonization)

Question

i need to implement Skeletonization in Emgu CV but i not have success. I have a code mentioned on the website below but it does not work : Skeletonization using EmguCV^]

This code below DONT work:

Image<Gray, Byte> eroded = new Image<Gray, byte>(img2.Size);
    Image<Gray, Byte> temp = new Image<Gray, byte>(img2.Size);
    Image<Gray, Byte> skel = new Image<Gray, byte>(img2.Size);
    skel.SetValue(0);
    CvInvoke.cvThreshold(img2, img2, 127, 256, 0);
    StructuringElementEx element = new StructuringElementEx(3, 3, 1, 1, Emgu.CV.CvEnum.CV_ELEMENT_SHAPE.CV_SHAPE_CROSS);
    bool done = false;

    while (!done)
    {
        CvInvoke.cvErode(img2, eroded, element,1);
        CvInvoke.cvDilate(eroded, temp, element,1);
        temp = img2.Sub(temp);
        skel = skel | temp;
        img2 = eroded;
        if (CvInvoke.cvCountNonZero(img2) == 0) done = true;
    }

This code WORK but is very slow in video (sequential frames)

Image<Gray, byte> Skeleton(Image<Gray, byte> orgImg)
    {
        Image<Gray, byte> skel = new Image<Gray, byte>(orgImg.Size);
        for (int y = 0; y < skel.Height; y++)
            for (int x = 0; x < skel.Width; x++)
                skel.Data[y, x, 0] = 0;

        imageBoxOutputROI.Image = skel;

        Image<Gray, byte> img = skel.Copy();
        for (int y = 0; y < skel.Height; y++)
            for (int x = 0; x < skel.Width; x++)
                img.Data[y, x, 0] = orgImg.Data[y, x, 0];

        StructuringElementEx element;
        element = new StructuringElementEx(3, 3, 1, 1, Emgu.CV.CvEnum.CV_ELEMENT_SHAPE.CV_SHAPE_CROSS);
        Image<Gray, byte> temp;

        bool done = false;
        do
        {
            temp = img.MorphologyEx(element, Emgu.CV.CvEnum.CV_MORPH_OP.CV_MOP_OPEN, 1);
            temp = temp.Not();
            temp = temp.And(img);
            skel = skel.Or(temp);
            img = img.Erode(1);
            double[] min, max;
            Point[] pmin, pmax;
            img.MinMax(out min, out max, out pmin, out pmax);
            done = (max[0] == 0);
        } while (!done);

        return skel;
    }

Input image:

I need a help to implement the skeletonization code.

A research in below sites but i not hace success: http://felix.abecassis.me/2011/09/opencv-morphological-skeleton/[^]

https://stackoverflow.com/questions/26850944/skeleton-of-an-image-in-emgucv[^]

https://stackoverflow.com/questions/26850944/skeleton-of-an-image-in-emgucv[^]

I am grateful for any help .

Richard J. Algarve

Answer 1

Try below code.

It works for me (My input image has white background). I think the problem with your code is Sub and Or operators.

    public static Bitmap Skelatanize(Bitmap image)
    {
        Image<Gray, byte> imgOld = new Image<Gray, byte>(image);
        Image<Gray, byte> img2 = (new Image<Gray, byte>(imgOld.Width, imgOld.Height, new Gray(255))).Sub(imgOld);
        Image<Gray, byte> eroded = new Image<Gray, byte>(img2.Size);
        Image<Gray, byte> temp = new Image<Gray, byte>(img2.Size);
        Image<Gray, byte> skel = new Image<Gray, byte>(img2.Size);
        skel.SetValue(0);
        CvInvoke.Threshold(img2, img2, 127, 256, 0);
        var element = CvInvoke.GetStructuringElement(ElementShape.Cross, new Size(3, 3), new Point(-1, -1));
        bool done = false;

        while (!done)
        {
            CvInvoke.Erode(img2, eroded, element, new Point(-1, -1), 1, BorderType.Reflect, default(MCvScalar));
            CvInvoke.Dilate(eroded, temp, element, new Point(-1, -1), 1, BorderType.Reflect, default(MCvScalar));
            CvInvoke.Subtract(img2, temp, temp);
            CvInvoke.BitwiseOr(skel, temp, skel);
            eroded.CopyTo(img2);
            if (CvInvoke.CountNonZero(img2) == 0) done = true;
        }
        return skel.Bitmap;
    }