private void buttonTrain_Click(object sender, EventArgs e)
{
if (radioButtonAll.Checked)
{
ReadData.readCompressedData();
}
else
{
ReadData.readSmallCompressedData();
}
RadialBasisNet RB = new RadialBasisNet();
char[] delimiterChars = { ' ' };
string[] words = textBoxLayers.Text.Split(delimiterChars);
List <int> layersList = new List <int>();
layersList.Add(256);
for (int i = 0; i < words.Length; i++)
{
if (words[i] != "")
{
layersList.Add(Convert.ToInt32(words[i]));
}
}
layersList.Add(3);
int[] layers = layersList.ToArray();
Neural_Network NN = new Neural_Network();
Tuple <double, double[, ]> t = NN.train(layers, Convert.ToInt32(textBoxNum_Iteraions.Text), Convert.ToDouble(textBoxEta.Text));
ConfusionMatrix CM = new ConfusionMatrix();
CM.fillMatrix(t.Item2);
textBoxAccuracy.Text = t.Item1.ToString() + '%';
CM.Show();
}
private void buttonClassify_Click(object sender, EventArgs e)
{
if (image != null)
{
Neural_Network pretrainedNN = new Neural_Network();
image = new Bitmap(image, 50, 50);
double[,] input = ImageProcessing.imageToGrayscale(image);
ReadData.normalizeImage(ref input);
PCA pca = new PCA();
pca.readPretrainedWeights();
input = Utilities.addBias(pca.compress(input));
textBoxClassification.Text = pretrainedNN.classify(input);
}
else
{
MessageBox.Show("Enter picture first");
}
}