private void CalculateCorrelations()
{
StatusBarProgressBar.Visible = true;
if (!CovarianceWorker.IsBusy)
{
CovarianceWorker.RunWorkerAsync();
}
Cursor.Current = Cursors.WaitCursor;
}
private void CovarianceTimer_Tick(object sender, EventArgs e)
{
if (dataIsInvalid)
{
if (CovarianceWorker.IsBusy)
{
CovarianceWorker.CancelAsync();
}
sampleSize = (int)CorrelationSampleSize.Value;
CalculateCorrelations();
dataIsInvalid = false;
}
}
private void CovarianceWorker_DoWork(object sender, DoWorkEventArgs e)
{
int progress = 0;
int maxSteps = 3 * sampleSize;
int step;
x = new int [sampleSize];
y = new int [sampleSize];
xy = new int [sampleSize];
xy_hor = new int [sampleSize];
xy_ver = new int [sampleSize];
Random rand = new Random();
step = 1;
for (int i = 0; i < sampleSize && !e.Cancel; ++i)
{
x[i] = rand.Next(0, input.Width - 1);
y[i] = rand.Next(0, input.Height - 1);
xy[i] = ArgbToY(input.GetPixel(x[i], y[i]));
if (x[i] > 0)
{
xy_hor[i] = ArgbToY(input.GetPixel(x[i] - 1, y[i]));
}
else
{
xy_hor[i] = ArgbToY(input.GetPixel(x[i] + 1, y[i]));
}
if (y[i] > 0)
{
xy_ver[i] = ArgbToY(input.GetPixel(x[i], y[i] - 1));
}
else
{
xy_ver[i] = ArgbToY(input.GetPixel(x[i], y[i] + 1));
}
if (i % 10 == 0)
{
progress = (i * step * 100) / maxSteps;
CovarianceWorker.ReportProgress(progress);
}
}
if (e.Cancel)
{
return;
}
step = 2;
expectedValueXY = 0;
expectedValueXYHor = 0;
expectedValueXYVer = 0;
for (int i = 0; i < sampleSize && !e.Cancel; ++i)
{
expectedValueXY += (double)xy[i];
expectedValueXYHor += (double)xy_hor[i];
expectedValueXYVer += (double)xy_ver[i];
if (i % 10 == 0)
{
progress = (i * step * 100) / maxSteps;
CovarianceWorker.ReportProgress(progress);
}
}
if (e.Cancel)
{
return;
}
expectedValueXY /= (double)sampleSize;
expectedValueXYHor /= (double)sampleSize;
expectedValueXYVer /= (double)sampleSize;
step = 3;
varianceXY = 0;
varianceXYHor = 0;
varianceXYVer = 0;
covarianceXYHor = 0;
covarianceXYVer = 0;
for (int i = 0; i < sampleSize && !e.Cancel; ++i)
{
double residual = (double)xy[i] - expectedValueXY;
double residualHor = (double)xy_hor[i] - expectedValueXY;
double residualVer = (double)xy_ver[i] - expectedValueXY;
varianceXY += Math.Pow(residual, 2.0);
varianceXYHor += Math.Pow(residualHor, 2.0);
varianceXYVer += Math.Pow(residualVer, 2.0);
covarianceXYHor += (residual * residualHor);
covarianceXYVer += (residual * residualVer);
if (i % 10 == 0)
{
progress = (i * step * 100) / maxSteps;
CovarianceWorker.ReportProgress(progress);
}
}
if (e.Cancel)
{
return;
}
varianceXY /= (double)sampleSize;
varianceXYHor /= (double)sampleSize;
varianceXYVer /= (double)sampleSize;
covarianceXYHor /= (double)sampleSize;
covarianceXYVer /= (double)sampleSize;
correlationCoeffHor = covarianceXYHor
/ (Math.Sqrt(varianceXY) * Math.Sqrt(varianceXYHor));
correlationCoeffVer = covarianceXYVer
/ (Math.Sqrt(varianceXY) * Math.Sqrt(varianceXYVer));
} // End method CovarianceWorker_DoWork().
