public void Train(byte[,,] imageData)
{
int trainSamples = 0;
FeatureComputer featureComputer = new FeatureComputer(imageData);
foreach (List <Point> currentTagPoints in this.tagPoints.Values)
{
trainSamples += currentTagPoints.Count;
}
////if (trainSamples > 0)
////{
//// double[][] inputs = new double[trainSamples][];
//// int[] outputs = new int[trainSamples];
//// this.ClearModels();
//// List<string> trueResponses = new List<string>();
//// List<DecisionVariable> attributes = new List<DecisionVariable>(FeatureComputer.NumberOfFeatures);
//// for (int i = 0; i < FeatureComputer.NumberOfFeatures; i++)
//// {
//// string columnName = "Feature" + i.ToString(CultureInfo.InvariantCulture);
//// attributes.Add(new DecisionVariable(columnName, new Accord.DoubleRange(double.MinValue, double.MaxValue)));
//// }
//// int trainSampleIndex = 0;
//// // Fill the train matrix
//// foreach (string label in this.tagPoints.Keys)
//// {
//// foreach (Point tagPoint in this.tagPoints[label])
//// {
//// float[] trainingData = featureComputer.ComputeFeatures(tagPoint);
//// inputs[trainSampleIndex] = new double[FeatureComputer.NumberOfFeatures];
//// Array.Copy(trainingData, inputs[trainSampleIndex], FeatureComputer.NumberOfFeatures);
//// trueResponses.Add(label);
//// trainSampleIndex++;
//// }
//// }
//// // Train the models
//// foreach (string label in this.tagPoints.Keys)
//// {
//// int responseIndex = 0;
//// // Prepare the responses matrix
//// foreach (string responseLabel in trueResponses)
//// {
//// outputs[responseIndex] = (responseLabel == label) ? 1 : 0;
//// responseIndex++;
//// }
//// DecisionTree decisionTree = new DecisionTree(attributes, this.tagPoints.Count);
//// C45Learning c45Learning = new C45Learning(decisionTree);
//// c45Learning.Learn(inputs, outputs);
//// this.Models.Add(label, decisionTree);
//// }
////}
}
