org.openimaj.math.matrix.algorithm.whitening

Class WhiteningTransform

java.lang.Object

org.openimaj.math.matrix.algorithm.whitening.WhiteningTransform

Direct Known Subclasses:

PCAWhitening

@Reference(type=Book,
           author={"Hyvrinen, Aapo","Hurri, Jarmo","Hoyer, Patrick O."},
           title="Natural Image Statistics: A Probabilistic Approach to Early Computational Vision.",
           year="2009",
           edition="1st",
           publisher="Springer Publishing Company, Incorporated",
           customData={"isbn","1848824904, 9781848824904"})
public abstract class WhiteningTransform
extends Object

Abstract base class for whitening transforms ("sphering"). Whitening transforms rescale data to remove correlations; more specifically the idea of whitening is that it completely removes the second-order information (correlations and (co)-variances) of the data.

Author:

Jonathon Hare (jsh2@ecs.soton.ac.uk)

Constructor Summary

Constructors

Constructor and Description
WhiteningTransform()

Method Summary

Modifier and Type	Method and Description
abstract void	train(double[][] data) Train the whitening transform with the given features.
abstract double[]	whiten(double[] vector) Apply the whitening transform to the given vector.
double[][]	whiten(double[][] vector) Apply the whitening transform to the given vectors.

Methods inherited from class java.lang.Object

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

Constructor Detail

WhiteningTransform

public WhiteningTransform()

Method Detail

whiten

public abstract double[] whiten(double[] vector)

Apply the whitening transform to the given vector.

Parameters:

vector - the vector

Returns:

the whitened vector

whiten

public double[][] whiten(double[][] vector)

Apply the whitening transform to the given vectors.

Parameters:

vector - the vectors; one per row

Returns:

the whitened vector

train

public abstract void train(double[][] data)

Train the whitening transform with the given features.

Parameters:

data - the data (one feature per row)