// Private members
#region Private Members
// Create Gaussian filter
private void CreateFilter( )
{
// create Gaussian function
AForge.Math.Gaussian gaus = new AForge.Math.Gaussian(sigma);
// create kernel
double[,] kernel = gaus.Kernel2D(size);
double min = kernel[0, 0];
// integer kernel
int[,] intKernel = new int[size, size];
int divisor = 0;
for (int i = 0; i < size; i++)
{
for (int j = 0; j < size; j++)
{
double v = kernel[i, j] / min;
if (v > ushort.MaxValue)
{
v = ushort.MaxValue;
}
intKernel[i, j] = (int)v;
// collect divisor
divisor += intKernel[i, j];
}
}
// update filter
this.Kernel = intKernel;
this.Divisor = divisor;
}
// Private members
#region Private Members
// Create Gaussian filter
private void CreateFilter( )
{
// create Gaussian function
var gaus = new AForge.Math.Gaussian(this.sigma);
// create kernel
var kernel = gaus.Kernel2D(this.size);
var min = kernel[0, 0];
// integer kernel
var intKernel = new int[this.size, this.size];
var divisor = 0;
for (var i = 0; i < this.size; i++)
{
for (var j = 0; j < this.size; j++)
{
var v = kernel[i, j] / min;
if (v > ushort.MaxValue)
{
v = ushort.MaxValue;
}
intKernel[i, j] = (int)v;
// collect divisor
divisor += intKernel[i, j];
}
}
// update filter
this.Kernel = intKernel;
this.Divisor = divisor;
}
// Private members
#region Private Members
// Create Gaussian filter
private void CreateFilter( )
{
// create Gaussian function
AForge.Math.Gaussian gaus = new AForge.Math.Gaussian(sigma);
// create kernel
double[,] kernel = gaus.Kernel2D(size);
double min = kernel[0, 0];
// integer kernel
int[,] intKernel = new int[size, size];
int sum = 0;
int divisor = 0;
// calculate integer kernel
for (int i = 0; i < size; i++)
{
for (int j = 0; j < size; j++)
{
double v = kernel[i, j] / min;
if (v > ushort.MaxValue)
{
v = ushort.MaxValue;
}
intKernel[i, j] = (int)v;
// collect sum
sum += intKernel[i, j];
}
}
// recalc kernel
int c = size >> 1;
for (int i = 0; i < size; i++)
{
for (int j = 0; j < size; j++)
{
if ((i == c) && (j == c))
{
// calculate central value
intKernel[i, j] = 2 * sum - intKernel[i, j];
}
else
{
// invert value
intKernel[i, j] = -intKernel[i, j];
}
// collect divisor
divisor += intKernel[i, j];
}
}
// update filter
this.Kernel = intKernel;
this.Divisor = divisor;
}
// Private members
#region Private Members
// Create Gaussian filter
private void CreateFilter()
{
// create Gaussian function
AForge.Math.Gaussian gaus = new AForge.Math.Gaussian(sigma);
// create kernel
int[,] kernel = gaus.KernelDiscret2D(size);
// create filter
filter = new Correlation(kernel);
}
// Private members
#region Private Members
// Create Gaussian filter
private void CreateFilter( )
{
// create Gaussian function
var gaus = new AForge.Math.Gaussian(this.sigma);
// create kernel
var kernel = gaus.Kernel2D(this.size);
var min = kernel[0, 0];
// integer kernel
var intKernel = new int[this.size, this.size];
var sum = 0;
var divisor = 0;
// calculate integer kernel
for (var i = 0; i < this.size; i++)
{
for (var j = 0; j < this.size; j++)
{
var v = kernel[i, j] / min;
if (v > ushort.MaxValue)
{
v = ushort.MaxValue;
}
intKernel[i, j] = (int)v;
// collect sum
sum += intKernel[i, j];
}
}
// recalc kernel
var c = this.size >> 1;
for (var i = 0; i < this.size; i++)
{
for (var j = 0; j < this.size; j++)
{
if ((i == c) && (j == c))
{
// calculate central value
intKernel[i, j] = 2 * sum - intKernel[i, j];
}
else
{
// invert value
intKernel[i, j] = -intKernel[i, j];
}
// collect divisor
divisor += intKernel[i, j];
}
}
// update filter
this.Kernel = intKernel;
this.Divisor = divisor;
}
// Private members
#region Private Members
// Create Gaussian filter
private void CreateFilter()
{
// create Gaussian function
AForge.Math.Gaussian gaus = new AForge.Math.Gaussian(sigma);
// create kernel
int[,] kernel = gaus.KernelDiscret2D(size);
// create filter
filter = new Correlation(kernel);
}
private void createSmoothingKernel()
{
AForge.Math.Gaussian g = new AForge.Math.Gaussian(smoothingSigma);
gaussianKernel = g.Kernel(smoothingKernelSize);
// Normalize
double norm = gaussianKernel.Euclidean();
gaussianKernel = gaussianKernel.Divide(norm);
}
/// <summary>
/// Smooths an image with Gaussian kernel.
/// </summary>
/// <param name="img">Input image.</param>
/// <param name="kernelSize">Kernel size.</param>
/// <param name="sigma">Sigma (standard deviation)</param>
/// <returns>Smoothed image.</returns>
public static TColor[,] SmoothGaussian <TColor>(this TColor[,] img, int kernelSize, double sigma)
where TColor : struct, IColor <float>
{
var gaussianKernel = new AForge.Math.Gaussian(sigma); //TODO: it would be nice if I could get separated kernel
var kernel = gaussianKernel.Kernel2D(kernelSize);
//normalize kernel
double factor = 1d / kernel.Sum().Sum();
kernel.ApplyInPlace(x => x * factor);
return(img.Convolve(kernel.ToSingle()));
}
// Private members
#region Private Members
// Create Gaussian filter
private void CreateFilter()
{
// create Gaussian function
AForge.Math.Gaussian gaus = new AForge.Math.Gaussian(sigma);
// create Gaussian kernel
int[,] kernel = gaus.KernelDiscret2D(size);
// calculte sum of the kernel
int sum = 0;
for (int i = 0; i < size; i++)
{
for (int j = 0; j < size; j++)
{
sum += kernel[i, j];
}
}
// recalc kernel
int c = size >> 1;
for (int i = 0; i < size; i++)
{
for (int j = 0; j < size; j++)
{
if ((i == c) && (j == c))
{
// calculate central value
kernel[i, j] = 2 * sum - kernel[i, j];
}
else
{
// invert value
kernel[i, j] = -kernel[i, j];
}
}
}
// create filter
filter = new Correlation(kernel);
}
// Private members
#region Private Members
// Create Gaussian filter
private void CreateFilter( )
{
// create Gaussian function
AForge.Math.Gaussian gaus = new AForge.Math.Gaussian( sigma );
// create kernel
double[,] kernel = gaus.Kernel2D( size );
double min = kernel[0, 0];
// integer kernel
int[,] intKernel = new int[size, size];
int divisor = 0;
for ( int i = 0; i < size; i++ )
{
for ( int j = 0; j < size; j++ )
{
double v = kernel[i, j] / min;
if ( v > ushort.MaxValue )
{
v = ushort.MaxValue;
}
intKernel[i, j] = (int) v;
// collect divisor
divisor += intKernel[i, j];
}
}
// update filter
this.Kernel = intKernel;
this.Divisor = divisor;
}
private void createGaussian()
{
double[] aforgeKernel = new AForge.Math.Gaussian(sigma).Kernel(size);
this.kernel = Array.ConvertAll<double, float>(aforgeKernel, Convert.ToSingle);
}
// Private members
#region Private Members
// Create Gaussian filter
private void CreateFilter( )
{
// create Gaussian function
AForge.Math.Gaussian gaus = new AForge.Math.Gaussian( sigma );
// create kernel
double[,] kernel = gaus.Kernel2D( size );
double min = kernel[0, 0];
// integer kernel
int[,] intKernel = new int[size, size];
int sum = 0;
int divisor = 0;
// calculate integer kernel
for ( int i = 0; i < size; i++ )
{
for ( int j = 0; j < size; j++ )
{
double v = kernel[i, j] / min;
if ( v > ushort.MaxValue )
{
v = ushort.MaxValue;
}
intKernel[i, j] = (int) v;
// collect sum
sum += intKernel[i, j];
}
}
// recalc kernel
int c = size >> 1;
for ( int i = 0; i < size; i++ )
{
for ( int j = 0; j < size; j++ )
{
if ( ( i == c ) && ( j == c ) )
{
// calculate central value
intKernel[i, j] = 2 * sum - intKernel[i, j];
}
else
{
// invert value
intKernel[i, j] = -intKernel[i, j];
}
// collect divisor
divisor += intKernel[i, j];
}
}
// update filter
this.Kernel = intKernel;
this.Divisor = divisor;
}
private void createGaussian()
{
double[] aforgeKernel = new AForge.Math.Gaussian(sigma).Kernel(size);
this.kernel = Array.ConvertAll <double, float>(aforgeKernel, Convert.ToSingle);
}
// Create Gaussian filter
private void CreateFilter()
{
// create Gaussian function
AForge.Math.Gaussian gaus = new AForge.Math.Gaussian(sigma);
// create Gaussian kernel
int[,] kernel = gaus.KernelDiscret2D(size);
// calculte sum of the kernel
int sum = 0;
for (int i = 0; i < size; i++)
{
for (int j = 0; j < size; j++)
{
sum += kernel[i, j];
}
}
// recalc kernel
int c = size >> 1;
for (int i = 0; i < size; i++)
{
for (int j = 0; j < size; j++)
{
if ((i == c) && (j == c))
{
// calculate central value
kernel[i, j] = 2 * sum - kernel[i, j];
}
else
{
// invert value
kernel[i, j] = -kernel[i, j];
}
}
}
// create filter
filter = new Correlation(kernel);
}
