org.openimaj.image.processing.algorithm

Class AnisotropicDiffusion

java.lang.Object

org.openimaj.image.processing.algorithm.AnisotropicDiffusion

All Implemented Interfaces:

ImageProcessor<FImage>, Processor<FImage>, SinglebandImageProcessor<Float,FImage>

@Reference(type=Article,
           author={"Perona, P.","Malik, J."},
           title="Scale-Space and Edge Detection Using Anisotropic Diffusion",
           year="1990",
           journal="IEEE Trans. Pattern Anal. Mach. Intell.",
           pages={"629","","639"},
           url="http://dx.doi.org/10.1109/34.56205",
           month="July",
           number="7",
           publisher="IEEE Computer Society",
           volume="12",
           customData={"issue_date","July 1990","issn","0162-8828","numpages","11","doi","10.1109/34.56205","acmid","78304","address","Washington, DC, USA"})
public class AnisotropicDiffusion
extends Object
implements SinglebandImageProcessor<Float,FImage>

Implementation of Perona & Malik's image filtering by anisotropic diffusion. Enables edge-preserving image smoothing.

Author:

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

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static interface	AnisotropicDiffusion.ConductionCoefficientFunction Interface describing a function that computes the conduction coefficient as a function of space and gradient magnitude.
static class	AnisotropicDiffusion.HighConstrastEdgeFunction The first conduction function proposed by Perona & Malik, that privileges high contrast edges over low constrast ones.
static class	AnisotropicDiffusion.WideRegionFunction The second conduction function proposed by Perona & Malik, that privileges wider regions over smaller ones.

Nested classes/interfaces inherited from interface org.openimaj.image.processor.SinglebandImageProcessor

SinglebandImageProcessor.Processable<T,S extends Image<T,S>,I extends Image<?,I>>

Field Summary

Fields

Modifier and Type	Field and Description
protected AnisotropicDiffusion.ConductionCoefficientFunction	function
protected float	lambda
protected ConnectedComponent.ConnectMode	neighbourMode
protected int	numIterations

Constructor Summary

Constructors

Constructor and Description
AnisotropicDiffusion(int numIterations, float lambda, AnisotropicDiffusion.ConductionCoefficientFunction function) Construct using the given parameters.
AnisotropicDiffusion(int numIterations, float lambda, AnisotropicDiffusion.ConductionCoefficientFunction function, ConnectedComponent.ConnectMode neighbourMode) Construct using the given parameters.

Method Summary

Modifier and Type	Method and Description
void	processImage(FImage image) Process an image.
void	processImageOneIteration(FImage image) Perform a single iteration of anisotropic diffusion

Methods inherited from class java.lang.Object

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

Field Detail

numIterations

protected int numIterations

lambda

protected float lambda

function

protected AnisotropicDiffusion.ConductionCoefficientFunction function

neighbourMode

protected ConnectedComponent.ConnectMode neighbourMode

Constructor Detail

AnisotropicDiffusion

public AnisotropicDiffusion(int numIterations,
                            float lambda,
                            AnisotropicDiffusion.ConductionCoefficientFunction function,
                            ConnectedComponent.ConnectMode neighbourMode)

Construct using the given parameters.

Parameters:

numIterations - number of iterations of filtering to apply

lambda - the integration constant (less than or equal to 1/4 for 4-neighbour mode and 1/8 for 8-neighbour)

function - the conduction coefficient function

neighbourMode - the neighbourhood mode

AnisotropicDiffusion

public AnisotropicDiffusion(int numIterations,
                            float lambda,
                            AnisotropicDiffusion.ConductionCoefficientFunction function)

Construct using the given parameters. 4-neighbour mode is used as per the original paper.

Parameters:

numIterations - number of iterations of filtering to apply

lambda - the integration constant (bigger than 0 and less than or equal to 1/4)

function - the conduction coefficient function

Method Detail

processImage

public void processImage(FImage image)

Description copied from interface: ImageProcessor

Process an image. Implementing classes must alter the image passed in-place or assign the output to the input using Image.internalAssign(Image).

Specified by:

processImage in interface ImageProcessor<FImage>

Parameters:

image - The image to process in place.

processImageOneIteration

public void processImageOneIteration(FImage image)

Perform a single iteration of anisotropic diffusion

Parameters:

image - the image