I want to get the difference between two images

I use this source code:

 public Bitmap GetDifference(Bitmap bm1, Bitmap bm2)
        {

            if (bm1.Size != bm2.Size)
                throw new ArgumentException("exception");


            var resultImage = new Bitmap(bm1.Width, bm1.Height);

            using (Graphics g = Graphics.FromImage(resultImage))
                g.Clear(Color.Transparent);


            for (int w = 0; w < bm1.Width; w++)
            {
                for (int h = 0; h < bm1.Height; h++)
                {
                    var bm2Color = bm2.GetPixel(w, h);
                    if (IsColorsDifferent(bm1.GetPixel(w, h), bm2Color))
                    {
                        resultImage.SetPixel(w, h, bm2Color);                    
                    }                  
                }
            }

            return resultImage;
        }


        bool IsColorsDifferent(Color c1, Color c2)
        {
            return c1 != c2;
        }

This source works good for me, but I have following issues:

For exaple if I choose image with a resolution 480x320 my resultImage has the same resolution (and this is corect, because I set in when created new Bitmap), but my Images have transparent color and I want to get result images with a solid color only.

Let us say - if the solid result images (solid color pixel) in 480x320 have 100x100 field I should to get just this solid Bitmap with a resolution 100x100.

To put it simply I need to get only rectangle that bordered with a solid color and cutt of alpha chanel.

Thanks!

• c#
• image-processing
• bitmap

Please see the class that I wrote below:

class ProcessingImage
    {         
        public Bitmap GetDifference(Bitmap bm1, Bitmap bm2)
        {            
            if (bm1.Size != bm2.Size)
                throw new ArgumentException("Images must have one size");

            var resultImage = new Bitmap(bm1.Width, bm1.Height);

            using (Graphics g = Graphics.FromImage(resultImage))
                g.Clear(Color.Transparent);

            int min_height = int.MaxValue;
            int min_width = int.MaxValue;
            int max_height = -1;
            int max_width = -1;

            for (int w = 0; w < bm1.Width; w++)
                for (int h = 0; h < bm1.Height; h++)
                {
                    var bm2Color = bm2.GetPixel(w, h);
                    if (IsColorsDifferent(bm1.GetPixel(w, h), bm2Color))
                    {
                        resultImage.SetPixel(w, h, bm2Color);
                        if (h < min_height) min_height = h;
                        if (w < min_width) min_width = w;
                        if (h > max_height) max_height = h;
                        if (w > max_width) max_width = w;
                    }                    
                }

            max_height = max_height + 1;    //Needed for get valid max height point, otherwise we lost one pixel.
            max_width = max_width + 1;      //Needed for get valid max width point, otherwise we lost one pixel.


            // Calculate original size for image without alpha chanel.

            int size_h = max_height - min_height;
            int size_w = max_width - min_width;

            var resizeImage = new Bitmap(size_w, size_h); // Creaete bitmap with new size.

            for (int w = 0; w < resultImage.Width; w++)
                for (int h = 0; h < resultImage.Height; h++)
                {
                    var color = resultImage.GetPixel(w, h);
                    if (color.A != 0)
                    {
                        // Move each pixels at a distance calculate before.
                        int x = w - min_width;
                        int y = h - min_height;
                        resizeImage.SetPixel(x, y, color);
                    }
                }           

            return resizeImage;
        }

        bool IsColorsDifferent(Color c1, Color c2)
        {
            return c1 != c2;
        }
    }

I think, this class can be modified to optimize, but now it works good. If you need return image with original size and without alpha channel - this is one of many solutions how to do it.