bijnum
Class BIJgifa

java.lang.Object
  |
  +--bijnum.BIJpca
        |
        +--bijnum.BIJgifa

public class BIJgifa

extends BIJpca

Implementation of GIFA (generalized indicator function approach). Does a maximization of images with maximum signal power and signal to noise ratio. Is a subset of BIJpca because it needs the PCA routines and output. (c) 2003 Michael Abramoff. All rights reserved.

Field Summary

float[]	gamma
int	k
float[][]	phi
float[][]	rho
float[]	snr

Fields inherited from class bijnum.BIJpca

a, an, eigenimages, eigenvalues, eigenvectors, totalVariance

Constructor Summary

BIJgifa()

Method Summary

float[][]	compute(float[][] a, int k) Compute GIFA for k conditions.
protected float[][]	submean(float[][] F, int k) Computes the mean of the submatrices of a MxN matrix F.

Methods inherited from class bijnum.BIJpca

compute, computeEigenimages, computeLinearCombination, computeLinearCombination, computeProjection, getCoordinates, getCoordinatesImage, getEigenImage, getTotalVariance, varianceSpectrum

Methods inherited from class java.lang.Object

, clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

phi

public float[][] phi

rho

public float[][] rho

gamma

public float[] gamma

snr

public float[] snr

k

public int k

Constructor Detail

BIJgifa

public BIJgifa()

Method Detail

compute

public float[][] compute(float[][] a,
                         int k)

Compute GIFA for k conditions.

submean

protected float[][] submean(float[][] F,
                            int k)

Computes the mean of the submatrices of a MxN matrix F. F is divided up in k submatrices of equal size, F = [F1, F2, Fk] The -ith submatrix Fi is (M/k)xN

Parameters:

F - a MxN matrix of the above form.

k - = number of submatrices

Returns:

a kxN matrix of the form [mean(G1), mean(G2, mean(Fk)]