---

IcaNet Class Reference

#include <IcaNet.hpp>

Collaboration diagram for IcaNet:

[legend]List of all members.

---

Detailed Description

Single Sphered ICA Separator.

Author:

Bernard De Cuyper

Version:

0.05

Date:

7/12/2003

 
Purpose:        Define a sphered ICA single cell neural net to get sub-gaussian or super-gaussian IC component.

Paper:  "Simple Neural Models for Independent Component Analysis.", 
                Aapo Hyvarinen and Erkki Oja, 18/02/1997 , Report University of Technology, Helsinski.

                "One Unit Learning Rules for Independent Component Analysis.", 
                Aapo Hyvarinen and Erkki Oja, 1997 , Helsinski.
                

@ Copyrights: Bernard De Cuyper & Eddy Fraiha 2003, Eggs & Pictures. MIT/Open BSD copyright model.

Public Methods
	IcaNet (double rate, double ax, double bx, bool superType)
	IcaNet (double rate, double ax, double bx, double cx, bool superType)
virtual	~IcaNet ()
virtual RealVector *	getW ()
virtual void	init (int size)
virtual void	learning (RealVector *sample, int nmax=1)
virtual double	forward (RealVector *src)
virtual void	output ()
Protected Methods
virtual double	g (double u)
virtual double	gMin (double u)
virtual double	gPlus (double u)
virtual void	learn (RealVector *sample)
Private Attributes
bool	superGaussian
bool	combined
double	learnRate
double	a
double	b
double	c
RealVector *	w
double	wNorm2
RealVector *	dw

---

The documentation for this class was generated from the following files:

• IcaNet.hpp
• Copy of IcaNet.cpp
• IcaNet.cpp

Restoreinpaint sourceforge project `C++/Java Image Processing, Restoration, Inpainting Project'.

Bernard De Cuyper: Open Project Leader: Concept, design and development. Bernard De Cuyper & Eddy Fraiha 2002, 2003. Bernard De Cuyper 2004. Open and free, for friendly usage only. Modifications on Belgium ground of this piece of artistic work, by governement institutions or companies, must be notified to Bernard De Cuyper.