I had the same question this very day, my workaround was to take image1 and image2 converted to 256x256 or 128x128 both translated AND then generate an image3 with the difference between them, then scan image3 checking the differences and returning the difference amount, I found out that the LOWER difference amount in % more equal the images are and more likely for them to be equal. This way you can identify if images are equal even if they're differently sized. here is the code.
double CompareImages(Bitmap InputImage1, Bitmap InputImage2, int Tollerance)
{
Bitmap Image1 = new Bitmap(InputImage1, new Size(128, 128));
Bitmap Image2 = new Bitmap(InputImage2, new Size(128, 128));
int Image1Size = Image1.Width * Image1.Height;
int Image2Size = Image2.Width * Image2.Height;
Bitmap Image3;
if (Image1Size > Image2Size)
{
Image1 = new Bitmap(Image1, Image2.Size);
Image3 = new Bitmap(Image2.Width, Image2.Height);
}
else
{
Image1 = new Bitmap(Image1, Image2.Size);
Image3 = new Bitmap(Image2.Width, Image2.Height);
}
for (int x = 0; x < Image1.Width; x++)
{
for (int y = 0; y < Image1.Height; y++)
{
Color Color1 = Image1.GetPixel(x, y);
Color Color2 = Image2.GetPixel(x, y);
int r = Color1.R > Color2.R ? Color1.R - Color2.R : Color2.R - Color1.R;
int g = Color1.G > Color2.G ? Color1.G - Color2.G : Color2.G - Color1.G;
int b = Color1.B > Color2.B ? Color1.B - Color2.B : Color2.B - Color1.B;
Image3.SetPixel(x, y, Color.FromArgb(r,g,b));
}
}
int Difference = 0;
for (int x = 0; x < Image1.Width; x++)
{
for (int y = 0; y < Image1.Height; y++)
{
Color Color1 = Image3.GetPixel(x, y);
int Media = (Color1.R + Color1.G + Color1.B) / 3;
if (Media > Tollerance)
Difference++;
}
}
double UsedSize = Image1Size > Image2Size ? Image2Size : Image1Size;
double result = Difference*100/UsedSize;
return Difference*100/UsedSize;
}
Tested here with over 900 images and it works like a charm x)
