Navigation

Related Articles

Back to Latest Articles

How To Make Thinning In Image Processing Work With C#

Thinning is a morphological operation in image processing, which eats away at the foregorund pixels and leaves only lines shaping objects.


Andraz Krzisnik
How To Make Thinning In Image Processing...

Thinning is a morphological operation in image processing, which removes foreground pixels from objects in the images. Furthermore, what it leaves behind are lines that still resemble the objects.

Morphological processes are non-linear operations, which means that their results depend on the shapes. So far, we worked only with binary images in morphology and we’re going to do so in this tutorial as well.

Binary images carry only two different intensities, which produce pixels with either black or white color. In morphological terms, we’ll also call them foreground and background pixels.

When we process images with morphological operations, we divide pixels into sets. Therefore, we end up with two sets and we apply these operations across these whole sets. However, there are exceptions where we limit operations to certain objects, like hole filling operation.

How does thinning work?

In order to eat away at foreground pixels that represent obejcts in the image, we’ll need to apply hit or miss transform elements. But, the objects have different shapes of edges around it, so we will need to use multiple structuring elements to address this problem.

Structuring elements can also be viewed as kernels in convolution process. But they are not the same, since convolution is a linear operation, which involves multiplying and summing.

These structuring elements we’re going to use will detect if foreground and background values overlap. And once they do, we will set the pixel that overlaps in the center to background color – black. However, we need to apply this process in a sequence.

We will use 8 different structuring elements that will detect edges in 8 different positions, one for every 45 degrees turn. But it won’t be enough if we apply this sequence only once. We need to go through this sequence multiple times, in order to reach convergence.

thinning process in image processing
Thinning process in image processing

Convergence simply represents, when we no longer apply anymore changes to the image. This way the results of thinning are usually just lines that briefly represent the objects.

Code

I’ve written a function that applies this operation to a binary image. However, the structuring elements that I used in it, are just two dimensional arrays of 1s and 0s. But in any case, I’ll leave a link to the whole demo project at the end of the post.

public static Bitmap Thinning(this Bitmap image)
     {
         int w = image.Width;
         int h = image.Height;

         BitmapData image_data = image.LockBits(
             new Rectangle(0, 0, w, h),
             ImageLockMode.ReadOnly,
             PixelFormat.Format8bppIndexed);

         int bytes = image_data.Stride * image_data.Height;
         byte[] buffer = new byte[bytes];

         Marshal.Copy(image_data.Scan0, buffer, 0, bytes);
         image.UnlockBits(image_data);

         int[][,] se =
         {
             Structuring_Element.left,
             Structuring_Element.topleft,
             Structuring_Element.top,
             Structuring_Element.topright,
             Structuring_Element.right,
             Structuring_Element.bottomright,
             Structuring_Element.bottom,
             Structuring_Element.bottomleft
         };

         int diff = 1;

         while (diff > 0)
         {
             diff = 0;
             for (int i = 0; i < se.Length; i++)
             {
                 for (int x = 1; x < w - 1; x++)
                 {
                     for (int y = 1; y < h - 1; y++)
                     {
                         int position = x + y * image_data.Stride;
                         List<bool> change = new List<bool>();
                         for (int kx = -1; kx < 2; kx++)
                         {
                             for (int ky = -1; ky < 2; ky++)
                             {
                                 int se_pos = position + kx + ky * image_data.Stride;
                                 int se_opp = position + (kx * -1) + (ky * -1) * image_data.Stride;
                                 if (se_pos == position)
                                 {
                                     continue;
                                 }

                                 if (buffer[se_pos] > 0 && se[i][kx+1,ky+1] == 1 && buffer[se_opp] == 0 && se[i][(kx* -1) + 1, (ky*-1) + 1] == 0)
                                 {
                                     change.Add(true);
                                 }
                             }
                         }

                         if (change.Count == 3 && buffer[position] > 0)
                         {
                             buffer[position] = 0;
                             diff++;
                         }
                     }
                 }
             }
         }

         Bitmap res_img = new Bitmap(w, h);
         BitmapData res_data = res_img.LockBits(
             new Rectangle(0, 0, w, h),
             ImageLockMode.WriteOnly,
             PixelFormat.Format8bppIndexed);
         Marshal.Copy(buffer, 0, res_data.Scan0, bytes);
         res_img.UnlockBits(res_data);

         return res_img;
     }

Conclusion

Thinning is a process, where I included elements from other morphological operations. It serves as a sor of stepping stone to a better understanding of processes I’ll cover in the future.

I hope this guide was helpful and I’m also including the link to the whole demo project, which you can try it out yourself. Just remember to use binary images with it.

Show Comments (0)

Comments

Related Articles

C# Tutorial

How To Use Butterworth Highpass Filter – C# Guide

Butterworth highpass filter is used to filter images in frequency domain. We can control how smooth transition beyond cut off frequency.

Posted on by Andraz Krzisnik
C# Image Processing

How To Use Butterworth Lowpass Filter – C# Guide

We use Butterworth lowpass filter to process images in the frequency domain. And by now, I assume you’re already familiar with Fourier transform and how to use it to...

Posted on by Andraz Krzisnik