public void openIMage()
{
Show_Image SI = new Show_Image();
SI.ShowDialog();
loadimage KM = new loadimage(SI.loodim);
Process currentProcess1 = System.Diagnostics.Process.GetCurrentProcess();
long totalBytesOfMemoryUsed2 = currentProcess1.WorkingSet64;
Benchmark.Start();
double[] res = this.S_V_M.testing(KM);
string output_to_Messagebox = "the output is ";
string[] f = this.RI.classes.ToArray();
Voting(KM.LM, res, null, KM.classes);
double maxf = res.Max();
//int indexk = 0;
//for (int d = 0; d < res.Length; d++)
//{
// if (res[d] >= maxf)
// indexk = d;
//}
//output_to_Messagebox = f[indexk];
for (int d = 0; d < res.Length; d++)
{
if (res[d] > 3)
{
output_to_Messagebox += " " + f[d];
}
}
Process currentProcess = System.Diagnostics.Process.GetCurrentProcess();
long totalBytesOfMemoryUsed = currentProcess.WorkingSet64;
Benchmark.End();
this.testingtime.Text = "The Testing Time : " + Benchmark.GetSeconds().ToString();
this.overall_accuracy_control.Text = this.overall_accuracy.ToString();
label2.Text = "Testing Memory is " + Math.Abs(totalBytesOfMemoryUsed - totalBytesOfMemoryUsed2) + " Bytes";
MessageBox.Show(output_to_Messagebox);
Showdrawing SD = new Showdrawing();
SD.IM.Image = pic_after_drow;
SD.Show();
}
public SVM_handler(ReadImages RM2, DataGridView confusion_matrix_control, TextBox overall_accuracy_control, Label trainingtime, Label testingtime, Label l1, Label L2)
{
RI = RM2;
confusion_matrix = new int[1, this.RI.classes.Count];
for (int f = 0; f < 1; f++)
{
for (int x = 0; x < this.RI.classes.Count; x++)
{
confusion_matrix[f, x] = 0;
}
}
this.confusion_matrix_control = confusion_matrix_control;
this.trainingtime = trainingtime;
this.testingtime = testingtime;
this.overall_accuracy_control = overall_accuracy_control;
Label1 = l1;
label2 = L2;
Process currentProcess = System.Diagnostics.Process.GetCurrentProcess();
long totalBytesOfMemoryUsed = currentProcess.WorkingSet64;
Benchmark.Start();
S_V_M = new Support_Vector_Machine(RI);
Benchmark.End();
this.trainingtime.Text = " The Training Time : " + Benchmark.GetSeconds().ToString();
Process currentProcess1 = System.Diagnostics.Process.GetCurrentProcess();
long totalBytesOfMemoryUsed1 = currentProcess1.WorkingSet64;
Label1.Text = "The Training Memmory is " + Math.Abs(totalBytesOfMemoryUsed - totalBytesOfMemoryUsed1) + " bytes";
for (int d = 0; d < this.RI.TestingImages.Length; ++d)
{
LM = new loadimage(RI.TestingImages[d]);
Voting(LM.LM, S_V_M.testing(LM), confusion_matrix, this.LM.classes);
}
display_results(this.confusion_matrix_control, this.overall_accuracy_control);
MessageBox.Show("Done Testing");
}
public double[] testing(loadimage InputSample)
{
double[] output;
float response = 0;
output = new double[this.DAtaset.classes.Count];
for (int trainindex = 0; trainindex < InputSample.image.Count; trainindex++)
{
for (int p1 = 0; p1 < DAtaset.TrainingSamples[0].Feature.Descriptor.Length; ++p1)
{
sample.Data[0, p1] = (InputSample.image[trainindex].Feature.Descriptor[p1]);
}
response = model.Predict(sample);
++output[Convert.ToInt32(response) - 1];
// mm[Convert.ToInt32(response) - 1].Add(InputSample.image[trainindex].Feature.KeyPoint.Point);
float x = InputSample.image[trainindex].Feature.KeyPoint.Point.X;
float y = InputSample.image[trainindex].Feature.KeyPoint.Point.Y;
float index = response;
mm[Convert.ToInt32(response) - 1].Add(new Point(Convert.ToInt32(x), Convert.ToInt32(y)));
}
return(output);
}
public void Run()
{
confusion_matrix = new int[1, this.ReadIm.classes.Count];
for (int f = 0; f < 1; f++)
{
for (int x = 0; x < this.ReadIm.classes.Count; x++)
{
confusion_matrix[f, x] = 0;
}
}
int[] w = GetNumberOfNeurins();
ML = new MultiLayerPerceptron(ReadIm, Eta, Epochs, NumberOfHiddenLayers, w);
Process currentProcess = System.Diagnostics.Process.GetCurrentProcess();
long totalBytesOfMemoryUsed = currentProcess.WorkingSet64;
Benchmark.Start();
ML.MLPTraining();
Benchmark.End();
Process currentProcess1 = System.Diagnostics.Process.GetCurrentProcess();
long totalBytesOfMemoryUsed2 = currentProcess1.WorkingSet64;
Label1.Text = "The Training Memory is " + Math.Abs(totalBytesOfMemoryUsed2 - totalBytesOfMemoryUsed) + " Bytes";
this.trainingtime.Text = "The Training time : " + Benchmark.GetSeconds().ToString();
double[] output = new double[ReadIm.classes.Count];
loadimage LMR;
for (int c = 0; c < this.ReadIm.TestingImages.Length; c++)
{
LMR = new loadimage(this.ReadIm.TestingImages[c]);
output = ML.MLPTesting(LMR);
Voting(LMR.LM, output, confusion_matrix, LMR.classes);
}
display_results(confusion_matrix_control, overall_accuracy_control);
overall_accuracy_control.Text = (((this.number_of_rigth_samples) / (double)(this.numberallsamplesinaccurcy)) * 100).ToString() + " %";
MessageBox.Show("Done Testing");
}
public void openimage()
{
Show_Image SI = new Show_Image();
SI.ShowDialog();
loadimage KM = new loadimage(SI.loodim);
Process currentProcess1 = System.Diagnostics.Process.GetCurrentProcess();
long totalBytesOfMemoryUsed2 = currentProcess1.WorkingSet64;
Benchmark.Start();
double[] res = this.ML.MLPTesting(KM);
string output_to_Messagebox = "the output is ";
string[] f = this.ReadIm.classes.ToArray();
Voting(KM.LM, res, null, KM.classes);
for (int d = 0; d < res.Length; d++)
{
if (res[d] > 30)
{
output_to_Messagebox += " " + f[d];
}
}
Benchmark.End();
Process currentProcess = System.Diagnostics.Process.GetCurrentProcess();
long totalBytesOfMemoryUsed = currentProcess.WorkingSet64;
Label2.Text = "The Testing Memory is " + Math.Abs(totalBytesOfMemoryUsed - totalBytesOfMemoryUsed2) + " Bytes";
this.testingtime.Text = "The Testing Time : " + Benchmark.GetSeconds().ToString();
MessageBox.Show(output_to_Messagebox);
Showdrawing SD = new Showdrawing();
SD.IM.Image = pic_after_drow;
SD.Show();
}
public double[] MLPTesting(loadimage InputSample)
{
double[] result = new double[this.Dataset.classes.Count];
double[] x = new double[AllLayers[AllLayers.Length - 1].Neurons.Length];
for (int g = 0; g < InputSample.image.Count; g++)
{
//for (int index = 0; index < InputSample.image[g].Feature.Descriptor.Length; index++)
//{
// InputSample.image[g].Feature.Descriptor[index] =
// (InputSample.image[g].Feature.Descriptor[index] - Dataset.Mean[index]) / Dataset.Max[index];
//}
AllLayers[0].SetY(InputSample.image[g].Feature.Descriptor);
///forward finished
for (int L = 1; L < AllLayers.Length; ++L)
{
AllLayers[L].Forward(AllLayers[L - 1]);
}
for (int i = 0; i < AllLayers[AllLayers.Length - 1].Neurons.Length; i++)
{
x[i] = AllLayers[AllLayers.Length - 1].Neurons[i].Y;
}
float x1 = InputSample.image[g].Feature.KeyPoint.Point.X;
float y1 = InputSample.image[g].Feature.KeyPoint.Point.Y;
int gs = (Convert.ToInt32(des(x)) - 1);
mm[gs].Add(new Point(Convert.ToInt32(x1), Convert.ToInt32(y1)));
result[des(x)]++;
}
return(result);
}
