public void AddDimension(double[] p)
{
double[] a = new double[1];
for (int i = 0; i < schwerpunkte.Length; ++i)
{
a[0] = Math.Pow(MatrixMafs.ArrMinusArr(p, schwerpunkte[i]).Average(), 2);
p = p.Concat(a).ToArray();
}
}
public static double[] GetVectorSchwerpunkt(double[][] p)
{
double[] x = new double[p[0].Length];
for (int i = 0; i < p.Length; ++i)
{
x = MatrixMafs.ArrAddArr(x, p[i]);
}
return(MatrixMafs.ArrDivNum(x, p.Length));
}
public void AddDimension()
{
double[] schwerpunkt = GetVectorSchwerpunkt(points.ToArray());
double[] a = new double[1];
for (int i = 0; i < points.Count; ++i)
{
a[0] = Math.Pow(MatrixMafs.ArrMinusArr(points[i], schwerpunkt).Average(), 2);
points[i] = points[i].Concat(a).ToArray();
}
}
public static Bitmap Convert2Bitmap(double[] arr0)
{
double[] arr = MatrixMafs.ArrMultNum(arr0, 255);
Bitmap bmp = new Bitmap(500, 500);
for (int i = 0; i < 250000; ++i)
{
bmp.SetPixel(i % 500, (i - (i % 500)) / 500, Color.FromArgb(Convert.ToByte(arr[i * 3]), Convert.ToByte(arr[i * 3 + 1]), Convert.ToByte(arr[i * 3 + 2])));
}
return(bmp);
}
public static double[][] AddDimension(double[][] p)
{
double[] schwerpunkt = GetVectorSchwerpunkt(p);
double[] a = new double[1];
for (int i = 0; i < p.Length; ++i)
{
a[0] = Math.Pow(MatrixMafs.ArrMinusArr(p[i], schwerpunkt).Average(), 2);
p[i] = p[i].Concat(a).ToArray();
}
return(p);
}
public void Train(int cycles, string animiePath, string echtBildPath)
{
string[] animieFiles = Directory.GetFiles(animiePath);
string[] echtBildFiles = Directory.GetFiles(echtBildPath);
for (int i = 0; i < cycles; ++i)
{
TrainingsCycle(MatrixMafs.ArrDivNum(GetDoubleArr(File.ReadAllBytes(echtBildFiles[i % echtBildFiles.Length])), 255), MatrixMafs.ArrDivNum(GetDoubleArr(File.ReadAllBytes(animieFiles[i % animieFiles.Length])), 255));
if ((i + 1) % savePer == 0)
{
File.Create(outputPath + i + ".arr");
Convert2Bitmap(decoder.GetOutput()).Save(outputPath + i + ".arr");
}
}
}
public void TrainDisciminator(int cycles, string animieFolder)
{
string[] files = Directory.GetFiles(animieFolder);
for (int i = 0; i < cycles; ++i)
{
discriminator.SetInput(MatrixMafs.ArrDivNum(GetDoubleArr(File.ReadAllBytes(files[i % files.Length])), 255));
discriminator.OutputBerechnen();
discriminator.DeltawertBerechen(new double[] { 1 });
discriminator.Backpropagation();
discriminator.SetInput(SimpleNet.GetRandArr(750000, 0, 1));
discriminator.OutputBerechnen();
discriminator.DeltawertBerechen(new double[] { 0 });
discriminator.Backpropagation();
if (i % 1000 == 0)
{
Console.WriteLine("Trainingscycle nr " + i);
}
}
}
public static double[] GaussisscherAlgorithmus(double[][] p)
{
double[][] x = p.Select(a => a.ToArray()).ToArray();
double[] y = new double[x.Length];
Parallel.For(0, y.Length, i =>
{
y[i] = -x[i][0];
});
Parallel.For(0, y.Length, i0 =>
{
for (int i1 = 0; i1 < x[0].Length - 1; ++i1)
{
x[i0][i1] = x[i0][i1 + 1];
}
x[i0][x[0].Length - 1] = 1;
});
for (int i = 0; i < x.Length; ++i)
{
if (x[i][i] == 0)
{
bool weiter = true;
for (int j = 0; j < x.Length && weiter; ++j)
{
if (x[j][i] != 0)
{
SwitchPosInArr(x, j, i);
}
}
}
}
for (int i0 = 0; i0 < x.Length; ++i0)
{
y[i0] /= x[i0][i0];
x[i0] = MatrixMafs.ArrDivNum(x[i0], x[i0][i0]);
Parallel.For(0, x.Length, i1 =>
{
if (i1 != i0)
{
y[i1] /= -x[i1][i0];
x[i1] = MatrixMafs.ArrDivNum(x[i1], -x[i1][i0]);
x[i1] = MatrixMafs.ArrAddArr(x[i1], x[i0]);
y[i1] += y[i0];
}
});
}
Parallel.For(0, x.Length, i =>
{
y[i] /= x[i][i];
x[i][i] = 1;
});
return(y);
}
