Main Page   Class Hierarchy   Compound List   File List   Compound Members   File Members  

Clustering.cpp File Reference

Gaussian mixture modeling of a multidimensionnal spaces by the MDL principle
FOR ANY EXTENSION, REQUIRES TO BE RE-IMPLEMENTED COMPLETLY. More...

#include <iostream>
#include <string>
#include <vector>
#include <fstream>
#include <cstdio>
#include <cstdlib>
#include <cmath>
#include <cassert>
#include <ctime>
#include "./TimeGaussMIX_MDL.hh"

Go to the source code of this file.

Functions

void usage (std::string argv0)
 Usage function.

int main (int argc, char *argv[])
 Main procedure for Gaussian mixture modeling. More...


Detailed Description

Gaussian mixture modeling of a multidimensionnal spaces by the MDL principle
FOR ANY EXTENSION, REQUIRES TO BE RE-IMPLEMENTED COMPLETLY.

This is the main program for infering by MDL using a data 2-part coding, a Gaussian mixture model of unknown complexity (paramters and number of Gaussians).
Modeling procedures can be chained. For clustering consecutive TL feature spaces, the initialization is done using the Gaussian parameters infered for the previous modeling. Furthermore, the algorithm checks that there is a sufficient number of Gaussians (fixed by the user) after an initialisation, and adds new Gaussian if necessary.
This is the main program which incorporates the users input/output parameters and launch the optimisation procedure by intializing and using methods of the object Temporal_MDL_classifier.

This algorithm, which makes the hypothesis that the Gaussian in the mixture are sufficiently separatyed to not interfer, may be extended to the general case.
However, to extend the modeling to the general case, the whole program seriously REQUIRES TO BE RE-CODED (This program was my first try in C++ so the developpment was chaotic!!)


Author:
P.Héas (IRIT / ENSAE / DLR / CNES)
Since:
January 2005

Definition in file Clustering.cpp.


Function Documentation

int main int    argc,
char *    argv[]
 

Main procedure for Gaussian mixture modeling.

The main procedure begins by a creating a Temporal_MDL_classifier object. After initializing this object, its method "Cluster_optimiser" is used to search the model which induces the minimum description length of the data. This method produces also the output classfile.

Parameters:
argv0  [1] Text file name of input multiband-images : the image bands of the different images are listed one after the other; The number of lines of this file must be (numberImagesXnumberBands)
argv0  [2] Text file name of output image classfiles
argv0  [3] Number of bands of each image
argv0  [4] Number of lines of the images
argv0  [5] Number of columns of the images
Optional parameters for selecting a spatial window :
Parameters:
argv0  [6] Column offset when reading the image file
argv0  [7] Line offset when reading the image file
argv0  [8] Number of columns of the image file
Optional parameters for spatial subsampling :
Parameters:
argv0  [9] factor to reduce each line (factor 1 = no subsampling)
argv0  [10] factor to reduce each column (factor 1 = no subsampling)
Optional parameters for Initialization :
Parameters:
argv0  [11] Minimum number of Gaussians in the mixture for initialization

Definition at line 71 of file Clustering.cpp.


Generated on Thu Feb 17 11:01:08 2005 for Gaussian Mixture Modeling by the MDL principle by doxygen1.2.14 written by Dimitri van Heesch, © 1997-2002