Package dev.brachtendorf.jimagehash.hashAlgorithms.filter

Class MaximumKernel

java.lang.Object

dev.brachtendorf.jimagehash.hashAlgorithms.filter.Kernel

dev.brachtendorf.jimagehash.hashAlgorithms.filter.NonAveragingKernel

dev.brachtendorf.jimagehash.hashAlgorithms.filter.MaximumKernel

All Implemented Interfaces:

Filter, Serializable

public class MaximumKernel
extends NonAveragingKernel

A maximum kernel is a non linear filter scanning the image and replacing every value with the maximum value found in the neighborhood. This maximum kernel allows a weight matrix to be supplied.

Example 1D kernel width 5 Kernel and no/uniform mask

        Values: 5 4 1 3 6
 

During convolution, the kernel looks at the value 1 and replaces it with the value 6 due to it being the maximum.

A weight mask [1 2 3 2 1] can give more emphasis on closer pixel values. An intermediary value matrix is calculated:

 Values * Mask = [5 8 3 6 6]
 

and it is found that the second value is the maximum. Now the unaltered vlaue at position 2 is taken. Therefore the 1 is replaced with the value 4.

Since:

2.0.0

Author:

Kilian

See Also:

MedianKernel, MinimumKernel, Serialized Form

Nested Class Summary

Nested classes/interfaces inherited from class dev.brachtendorf.jimagehash.hashAlgorithms.filter.Kernel

Kernel.EdgeHandlingStrategy, Kernel.GrayScaleFilter

Field Summary

Fields inherited from class dev.brachtendorf.jimagehash.hashAlgorithms.filter.Kernel

edgeHandling, mask

Constructor Summary

Constructors

Constructor	Description
MaximumKernel(double[][] mask)	Create a kernel with the given masks dimension.
MaximumKernel(int width, int height)	Create a maximum kernel with no weight matrix

Method Summary

Modifier and Type	Method	Description
protected double	calcValue(byte[][] input, int x, int y)
protected double	calcValue(double[][] input, int x, int y)
protected double	calcValue(int[][] input, int x, int y)
protected double	resolveMax(double[][] values)

Methods inherited from class dev.brachtendorf.jimagehash.hashAlgorithms.filter.NonAveragingKernel

computePotentialValues, computePotentialValues, computePotentialValues

Methods inherited from class dev.brachtendorf.jimagehash.hashAlgorithms.filter.Kernel

apply, apply, apply, applyByte, applyByte, applyInt, applyInt, boxFilter, boxFilterNormalized, boxFilterNormalizedSep, edgeDetectionFilter, embossHorizontontalFilter, embossLeftDiagonalFilter, embossleftRightFilter, embossRightDiagonalFilter, equals, filter, gaussianFilter, hashCode, identityFilter, toString, zeroFilter

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Constructor Detail

MaximumKernel

public MaximumKernel(int width,
                     int height)

Create a maximum kernel with no weight matrix

Parameters:

width - of the kernel

height - height of the kernel

MaximumKernel

public MaximumKernel(double[][] mask)

Create a kernel with the given masks dimension. The masks acts as weight filter increasing or decreasing the weight of the value during convolution. For an example see the javadoc of the class.

Parameters:

mask - weight matrix used to judge which value is the maximum

Method Detail

calcValue

protected double calcValue(byte[][] input,
                           int x,
                           int y)

Overrides:

calcValue in class Kernel

calcValue

protected double calcValue(int[][] input,
                           int x,
                           int y)

Overrides:

calcValue in class Kernel

calcValue

protected double calcValue(double[][] input,
                           int x,
                           int y)

Overrides:

calcValue in class Kernel

Parameters:

input - array

x - pixelToLookAt

y - pixelToLookAt

Returns:

convolutedPixel fo this x and y

resolveMax

protected double resolveMax(double[][] values)