/**
* Constructor of the FSM with result of the clustering, shape for temporal information calculation and tresholdMultiplier wihich
* indirectly affects recognition rate
*/
public FSM(KmeansClustering cl, Shape s, int tresholdMultuplier)
{
stateList = clusterToState(orderClusters(cl.ClusterList, s), tresholdMultuplier, s);
}
static void Main(string[] args)
{
int shapeNumber = 2; // shape to work with
//** Data for Clustering,FSM and Rcognition**
dataReader d = new dataReader("C:/Users/Daria/source/repos/SP_Bewegungserkennung/SP_Bewegungserkennung/KinectDaten_Pascal.csv");
List <Shape> shapes = d.readData();
d.scaleShapes(shapes);
//**Run Clustering**
KmeansClustering km = new KmeansClustering(shapes[shapeNumber], new point(25, 25), 2, Double.Epsilon);
km.clustering();
// **Visualize Clusters**
//visualiser vsl = new visualiser(km.CLlist);
//vsl.runVisualiser();
// **Create FSM**
FSM machine = new FSM(km, shapes[shapeNumber], 3);
// **Serialise FSM**
//FSM.serialize(machine, "testMachine.xml");
//FSM f2 = FSM.deserialize("testMachine.xml");
// **Visualize States**
List <point> pvis = new List <point>();
foreach (Gesture g in shapes[shapeNumber].getGestures())
{
foreach (point p in g.Points)
{
pvis.Add(p);
}
}
visualiser vsl2 = new visualiser(machine.stateList, pvis);
vsl2.runVisualiser();
// **Recognition**
foreach (Gesture g in shapes[shapeNumber].getGestures())
{
machine.recognize(g);
Console.WriteLine("not recognized");
}
Console.ReadLine();
return;
}