javax.media.jai
Class Histogram

java.lang.Object
  javax.media.jai.Histogram

All Implemented Interfaces:

Serializable

public class Histogram

extends Object

implements Serializable

This class represents a histogram accumulated from a RenderedImage.

A "bin" is a container, where each element stores the total number of pixel samples of an image whose values lie within a given range. A histogram of an image consists of a list of such bins whose range does not overlap with each other (mutually exclusive). For an image that has multiple samples per pixel (multi-banded images), a separate list of bins represents each individual band.

A "low-value" specifies the lowest inclusive pixel value to be checked, and a "high-value" specifies the highest exclusive pixel value to be checked. Therefore, the width of a bin (binWidth) is determined by (highValue - lowValue) / numberOfBins. The range of bin i is defined as from lowValue + i * binWidth inclusive to lowValue + (i + 1) * binWidth exclusive.

The image may have any data type. Its histogram may be accumulated over the entire image, or over a specific region-of-interest (ROI) within the image's bounds. Furthermore, the horizontal and vertical subsampling factors specify the rate of sampling in the two directions, so that only every nth pixel will be counted. This allows the accuracy of the histogram to be traded for the speed of the computation. Of course a subsampling rate of 1 means every pixel will be counted.

The "Histogram" operator generates the histogram data of an image and uses this object to store the final pixel counts. The operator returns an instance of this class when a request is made via the getProperty method for the "histogram" property. The actual bins may be obtained by calling the getBins method.

See Also:

ROI, HistogramDescriptor, Serialized Form

Constructor Summary

Histogram(int[] numBins, double[] lowValue, double[] highValue)
Constructor.

Histogram(int[] numBins, double[] lowValue, double[] highValue, int numBands)
Constructor.

Histogram(int numBins, double lowValue, double highValue, int numBands)
Constructor.

Method Summary

void	clearHistogram() Resets the values of all bins to zero.
void	countPixels(Raster raster, ROI roi, int xStart, int yStart, int xPeriod, int yPeriod) Accumulates the histogram of the pixels within a specific region-of-interest (ROI) by counting them based on the indicated horizontal and vertical sampling period.
double	getBinLowValue(int band, int bin) Returns the lowest inclusive pixel value of a given bin for a specific band.
int[][]	getBins() Returns the array of bands of bins, each bin is the histogram for a band, i.e., the format of the returned array is int[bands][bins].
int[]	getBins(int band) Returns the bins of the histogram for a specific band by reference.
int	getBinSize(int band, int bin) Returns the number of pixel samples found in a given bin for a specific band.
double[]	getEntropy() Returns the entropy of the histogram.
Histogram	getGaussianSmoothed(double standardDeviation) Computes a Gaussian smoothed version of the histogram.
double[]	getHighValue() Returns the highest exclusive pixel value checked for all bands.
double	getHighValue(int band) Returns the highest exclusive pixel value checked for a specific band.
double[]	getIterativeThreshold() Calculates the threshold using iterative bisection.
double[]	getLowValue() Returns the lowest inclusive pixel value checked for all bands.
double	getLowValue(int band) Returns the lowest inclusive pixel value checked for a specific band.
double[]	getMaxEntropyThreshold() Calculates the threshold which maximizes the entropy.
double[]	getMaxVarianceThreshold() Calculates the threshold which maximizes the ratio of the between-class variance to the within-class variance for each band.
double[]	getMean() Returns the mean values for all bands of the histogram.
double[]	getMinErrorThreshold() Calculates the threshold which minimizes the probability of error.
double[]	getMinFuzzinessThreshold() Calculates the threshold which minimizes the fuzziness.
double[]	getModeThreshold(double power) Calculates the threshold using the mode method.
double[]	getMoment(int moment, boolean isAbsolute, boolean isCentral) Returns a specified (absolute) (central) moment of the histogram.
int	getNumBands() Returns the number of bands of the histogram.
int[]	getNumBins() Returns the number of bins of the histogram for all bands.
int	getNumBins(int band) Returns the number of bins of the histogram for a specific band.
double[]	getPTileThreshold(double p) Calculates the p-tile threshold.
Histogram	getSmoothed(boolean isWeighted, int k) Computes a smoothed version of the histogram.
double[]	getStandardDeviation() Returns the standard deviation for all bands of the histogram.
int	getSubTotal(int band, int minBin, int maxBin) Returns the total bin count for the specified sub-range (via "min" and "max" bin) of the indicated band.
int[]	getTotals() Returns the total bin count over all bins for all bands.

Methods inherited from class java.lang.Object

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

Constructor Detail

Histogram

public Histogram(int[] numBins,
                 double[] lowValue,
                 double[] highValue)

Constructor.

This constructor should be used when numBins, lowValue, and/or highValues are different for each band of the image. The length of the arrays indicates the number of bands the image has, and the three arrays must have the same length.

Since this constructor has no way of knowing the actual number of bands of the image, the length of the arrays is not checked against anything. Therefore, it is very important that the caller supplies the correct arrays of correct lengths, or errors will occur.

Parameters:

numBins - The number of bins for each band of the image. The length of this array indicates the number of bands for this histogram.

lowValue - The lowest inclusive pixel value checked for each band.

highValue - The highest exclusive pixel value checked for each band.

Throws:

IllegalArgumentException - If numBins, lowValue, or highValue is null.

IllegalArgumentException - If the array length of the three arguments are not the same, or any array length is 0.

IllegalArgumentException - If the number of bins for any band is less than or equal to 0.

IllegalArgumentException - If the low-value of any band is greater than or equal to its corresponding high-value.

Histogram

public Histogram(int[] numBins,
                 double[] lowValue,
                 double[] highValue,
                 int numBands)

Constructor.

This constructor should be used when numBins, lowValue, and/or highValues may be different for each band of the image. The number of bands in the image is provided explicitly. If any of the arrays provided has a length which is less than the number of bands, the first element in that array is used to fill out the array to a length of numBands.

Parameters:

numBins - The number of bins for each band of the image.

lowValue - The lowest inclusive pixel value checked for each band.

highValue - The highest exclusive pixel value checked for each band.

numBands - The number of bands in the image.

Throws:

IllegalArgumentException - If numBins, lowValue, or highValue is null.

IllegalArgumentException - If any array length is 0.

IllegalArgumentException - If the number of bins for any band is less than or equal to 0.

IllegalArgumentException - If the low-value of any band is greater than or equal to its corresponding high-value.

IllegalArgumentException - If numBands is less than or equal to 0.

Since:

JAI 1.1

Histogram

public Histogram(int numBins,
                 double lowValue,
                 double highValue,
                 int numBands)

Constructor.

The same numBins, lowValue, and highValue is applied to every band of the image.

Parameters:

numBins - The number of bins for all bands of the image.

lowValue - The lowest inclusive pixel value checked for all bands.

highValue - The highest exclusive pixel value checked for all bands.

numBands - The number of bands of the image.

Throws:

IllegalArgumentException - If numBins or numBands is less than or equal to 0.

IllegalArgumentException - If lowValue is less than or equal to highValue.

Since:

JAI 1.1

Method Detail

getNumBins

public int[] getNumBins()

Returns the number of bins of the histogram for all bands.

getNumBins

public int getNumBins(int band)

Returns the number of bins of the histogram for a specific band.

Parameters:

band - The index of the band whose numBins is to be returned.

Throws:

ArrayIndexOutOfBoundsException - If an invalid band index is specified.

getLowValue

public double[] getLowValue()

Returns the lowest inclusive pixel value checked for all bands.

getLowValue

public double getLowValue(int band)

Returns the lowest inclusive pixel value checked for a specific band.

Parameters:

band - The index of the band whose lowValue is to be returned.

Throws:

ArrayIndexOutOfBoundsException - If an invalid band index is specified.

getHighValue

public double[] getHighValue()

Returns the highest exclusive pixel value checked for all bands.

getHighValue

public double getHighValue(int band)

Returns the highest exclusive pixel value checked for a specific band.

Parameters:

band - The index of the band whose highValue is to be returned.

Throws:

ArrayIndexOutOfBoundsException - If an invalid band index is specified.

getNumBands

public int getNumBands()

Returns the number of bands of the histogram. This value is the same as the number of bands of the bins, bins.length.

getBins

public int[][] getBins()

Returns the array of bands of bins, each bin is the histogram for a band, i.e., the format of the returned array is int[bands][bins].

getBins

public int[] getBins(int band)

Returns the bins of the histogram for a specific band by reference.

Parameters:

band - The index of the band whose bins are to be returned.

Throws:

ArrayIndexOutOfBoundsException - If an invalid band index is specified.

getBinSize

public int getBinSize(int band,
                      int bin)

Returns the number of pixel samples found in a given bin for a specific band.

Parameters:

band - The index of the band-of-interest.

bin - The index of the bin whose value is to be returned.

Throws:

ArrayIndexOutOfBoundsException - If an invalid band or bin index is specified.

getBinLowValue

public double getBinLowValue(int band,
                             int bin)

Returns the lowest inclusive pixel value of a given bin for a specific band.

Parameters:

band - The index of the band-of-interest.

bin - The index of the bin whose lowValue is to be returned.

Throws:

ArrayIndexOutOfBoundsException - If an invalid band index is specified.

clearHistogram

public void clearHistogram()

Resets the values of all bins to zero. If bins has not been initialized (null), this method does nothing.

getTotals

public int[] getTotals()

Returns the total bin count over all bins for all bands.

An array which stores the total bin count is kept in this class and a reference to this array is returned by this method for performance reasons. The elements of the returned array should not be modified or undefined errors may occur. The array format is int[numBands].

Since:

JAI 1.1

getSubTotal

public int getSubTotal(int band,
                       int minBin,
                       int maxBin)

Returns the total bin count for the specified sub-range (via "min" and "max" bin) of the indicated band. The sub-ragne must fall within the actual range of the bins.

Parameters:

band - The index of the band-of-interest.

minBin - The minimum bin index to be counted.

maxBin - The maximum bin index to be counted.

Throws:

ArrayIndexOutOfBoundsException - If an invalid band index is specified.

IllegalArgumentException - If minBin is greater than maxBin.

Since:

JAI 1.1

getMean

public double[] getMean()

Returns the mean values for all bands of the histogram.

Since:

JAI 1.1

countPixels

public void countPixels(Raster raster,
                        ROI roi,
                        int xStart,
                        int yStart,
                        int xPeriod,
                        int yPeriod)

Accumulates the histogram of the pixels within a specific region-of-interest (ROI) by counting them based on the indicated horizontal and vertical sampling period. The result is stored in the bins array and may be obtained by calling the getBins method.

The ROI specifies the region within which the pixels are counted. If it is null, the entire Raster is counted. If it is not null and does not intersect with the Raster, then this method returns without changing the bins.

The set of pixels to be counted may be obtained by intersecting the grid (xStart + i * xPeriod, yStart + j * yPeriod); i, j >= 0 with the ROI and the bounding rectangle of the Raster.

Parameters:

raster - The Raster that contains the pixels to be counted.

roi - The region-of-interest within which the pixels are counted.

xStart - The initial X sample coordinate.

yStart - The initial Y sample coordinate.

xPeriod - The X sampling period.

yPeriod - The Y sampling period.

Throws:

IllegalArgumentException - If raster is null.

IllegalArgumentException - If the pixels stored in the raster do not have the same number of bands (samples per pixel) as this histogram's bins.

getMoment

public double[] getMoment(int moment,
                          boolean isAbsolute,
                          boolean isCentral)

Returns a specified (absolute) (central) moment of the histogram.

The ith moment in each band is defined to be the mean of the image pixel values raised to the ith power in that band. For central moments the average of the ith power of the deviation from the mean is used. For absolute moments the absolute value of the exponentiated term is used.

Note that the mean is the first moment, the average energy the second central moment, etc.

Parameters:

moment - The moment number or index which must be positive or an IllegalArgumentException will be thrown.

isAbsolute - Whether to calculate the absolute moment.

isCentral - Whether to calculate the central moment.

Returns:

The requested (absolute) (central) moment of the histogram.

Since:

JAI 1.1

getStandardDeviation

public double[] getStandardDeviation()

Returns the standard deviation for all bands of the histogram. This is a convenience method as the returned values could be calculated using the first and second moments which are available via the moment generation function.

Returns:

The standard deviation values for all bands.

Since:

JAI 1.1

getEntropy

public double[] getEntropy()

Returns the entropy of the histogram.

The histogram entropy is defined to be the negation of the sum of the products of the probability associated with each bin with the base-2 log of the probability.

Returns:

The entropy of the histogram.

Since:

JAI 1.1

getSmoothed

public Histogram getSmoothed(boolean isWeighted,
                             int k)

Computes a smoothed version of the histogram.

Each band of the histogram is smoothed by averaging over a moving window of a size specified by the method parameter: if the value of the parameter is k then the width of the window is 2*k + 1. If the window runs off the end of the histogram only those values which intersect the histogram are taken into consideration. The smoothing may optionally be weighted to favor the central value using a "triangular" weighting. For example, for a value of k equal to 2 the central bin would have weight 1/3, the adjacent bins 2/9, and the next adjacent bins 1/9.

Parameters:

isWeighted - Whether bins will be weighted using a triangular weighting scheme favoring bins near the central bin.

k - The smoothing parameter which must be non-negative or an IllegalArgumentException will be thrown. If zero, the histogram object will be returned with no smoothing applied.

Returns:

A smoothed version of the histogram.

Since:

JAI 1.1

getGaussianSmoothed

public Histogram getGaussianSmoothed(double standardDeviation)

Computes a Gaussian smoothed version of the histogram.

Each band of the histogram is smoothed by discrete convolution with a kernel approximating a Gaussian impulse response with the specified standard deviation.

Parameters:

standardDeviation - The standard deviation of the Gaussian smoothing kernel which must be non-negative or an IllegalArgumentException will be thrown. If zero, the histogram object will be returned with no smoothing applied.

Returns:

A Gaussian smoothed version of the histogram.

Since:

JAI 1.1

getPTileThreshold

public double[] getPTileThreshold(double p)

Calculates the p-tile threshold.

Computes thresholds such that a specified proportion of the sample values in each band are below the threshold.

Parameters:

p - The proportion of samples in each band which should be below the threshold in the band. If p is not in the range (0.0, 1.0) an IllegalArgumentException will be thrown.

Returns:

The requested p-tile thresholds.

Since:

JAI 1.1

getModeThreshold

public double[] getModeThreshold(double power)

Calculates the threshold using the mode method.

The threshold is defined to be the minimum between two peaks. The first peak is the highest peak in the histogram. The second peak is the highest peak in the histogram weighted by a specified power of the distance from the first peak.

Parameters:

power - The exponent of the distance weighting from the first peak.

Returns:

The requested thresholds.

Since:

JAI 1.1

getIterativeThreshold

public double[] getIterativeThreshold()

Calculates the threshold using iterative bisection.

For each band an initial threshold is defined to be the midpoint of the range of data represented by the histogram. The mean value is calculated for each sub-histogram and a new threshold is defined as the arithmetic mean of the two sub-histogram means. This process is repeated until the threshold value no longer changes.

Returns:

The requested thresholds.

Since:

JAI 1.1

getMaxVarianceThreshold

public double[] getMaxVarianceThreshold()

Calculates the threshold which maximizes the ratio of the between-class variance to the within-class variance for each band.

Returns:

The requested thresholds.

Since:

JAI 1.1

getMaxEntropyThreshold

public double[] getMaxEntropyThreshold()

Calculates the threshold which maximizes the entropy.

The entropy of a range of gray levels is defined to be the negation of the sum of products of the probability and the logarithm thereof over all gray levels in the range. The maximum entropy threshold is defined to be that value which maximizes the sum of the entropy of the two ranges which are above and below the threshold, respectively. This computation is effected for each band.

Returns:

The requested thresholds.

Since:

JAI 1.1

getMinErrorThreshold

public double[] getMinErrorThreshold()

Calculates the threshold which minimizes the probability of error.

For each band the histogram is modeled as the sum of two Gaussian distributions and the threshold which minimizes the misclassification error is computed. If the underlying histogram is unimodal the mean value of each band will be returned as the threshold. The bimodality of the histogram for that band will be identically zero.

Returns:

The requested thresholds.

Since:

JAI 1.1

getMinFuzzinessThreshold

public double[] getMinFuzzinessThreshold()

Calculates the threshold which minimizes the fuzziness.

Since:

JAI 1.1