org.openimaj.image.model.pixel

Class SingleGaussianPixelModel

java.lang.Object

org.openimaj.image.model.pixel.PixelClassificationModel<Float[],MBFImage>

org.openimaj.image.model.pixel.MBFPixelClassificationModel

org.openimaj.image.model.pixel.SingleGaussianPixelModel

All Implemented Interfaces:

Serializable, Cloneable, ImageClassificationModel<MBFImage>, EstimatableModel<MBFImage,FImage>, Model<MBFImage,FImage>

public class SingleGaussianPixelModel
extends MBFPixelClassificationModel

An MBFPixelClassificationModel that classifies an individual pixel by comparing it to a CachingMultivariateGaussian. The Gaussian is learnt from the values of the positive pixel samples given in training. The probability returned by the classification is determined from the PDF of the Gaussian at the given pixel.

Author:

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

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected CachingMultivariateGaussian	gauss

Fields inherited from class org.openimaj.image.model.pixel.MBFPixelClassificationModel

ndims

Constructor Summary

Constructors

Constructor and Description
SingleGaussianPixelModel(int ndims) Construct with the given number of dimensions.

Method Summary

Modifier and Type	Method and Description
protected float	classifyPixel(Float[] pix)
SingleGaussianPixelModel	clone() Clone the model
void	learnModel(MBFImage... images) Learn the model from the given MBFImages.

Methods inherited from class org.openimaj.image.model.pixel.MBFPixelClassificationModel

getArray

Methods inherited from class org.openimaj.image.model.pixel.PixelClassificationModel

classifyImage, estimate, numItemsToEstimate, predict

Methods inherited from class java.lang.Object

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

Field Detail

gauss

protected CachingMultivariateGaussian gauss

Constructor Detail

SingleGaussianPixelModel

public SingleGaussianPixelModel(int ndims)

Construct with the given number of dimensions. This should be equal to the number of bands in the MBFImages you wish to classify.

Parameters:

ndims - the number of dimensions.

Method Detail

classifyPixel

protected float classifyPixel(Float[] pix)

Specified by:

classifyPixel in class PixelClassificationModel<Float[],MBFImage>

learnModel

public void learnModel(MBFImage... images)

Description copied from interface: ImageClassificationModel

Learn the model from the given MBFImages.

Parameters:

images - the images to learn from

clone

public SingleGaussianPixelModel clone()

Description copied from interface: EstimatableModel

Clone the model

Specified by:

clone in interface ImageClassificationModel<MBFImage>

Specified by:

clone in interface EstimatableModel<MBFImage,FImage>

Specified by:

clone in class PixelClassificationModel<Float[],MBFImage>

Returns:

a cloned copy