net.imglib2.algorithm.dog

Class DogDetection<T extends RealType<T> & NativeType<T>>

java.lang.Object

net.imglib2.algorithm.dog.DogDetection<T>

public class DogDetection<T extends RealType<T> & NativeType<T>>
extends Object

Author:

Stephan Preibisch, Tobias Pietzsch

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	DogDetection.ExtremaType
class	DogDetection.TypedDogDetection<F extends RealType<F> & NativeType<F>>

Field Summary

Fields

Modifier and Type	Field and Description
protected DogDetection.ExtremaType	extremaType
protected double	imageSigma
protected RandomAccessible<T>	input
protected Interval	interval
protected boolean	keepDoGImg
protected double	minf
protected double	minPeakValue
protected boolean	normalizeMinPeakValue
protected int	numThreads
protected double[]	pixelSize
protected double	sigmaLarger
protected double	sigmaSmaller
protected DogDetection.TypedDogDetection<?>	typedDogDetection

Constructor Summary

Constructors

Constructor and Description
DogDetection(RandomAccessible<T> input, Interval interval, double[] calibration, double sigmaSmaller, double sigmaLarger, DogDetection.ExtremaType extremaType, double minPeakValue, boolean normalizeMinPeakValue) Sets up a DogDetection with the specified parameters (does not do any computation yet).
DogDetection(RandomAccessible<T> input, Interval interval, double[] calibration, double sigmaSmaller, double sigmaLarger, DogDetection.ExtremaType extremaType, double minPeakValue, boolean normalizeMinPeakValue, F computationType) Sets up a DogDetection with the specified parameters (does not do any computation yet).
DogDetection(RandomAccessibleInterval<T> input, double[] calibration, double sigmaSmaller, double sigmaLarger, DogDetection.ExtremaType extremaType, double minPeakValue, boolean normalizeMinPeakValue)

Method Summary

Modifier and Type	Method and Description
double	getImageSigma()
boolean	getKeepDoGImg()
double	getMinf()
double	getMinPeakValue()
boolean	getNormalizeMinPeakValue()
int	getNumThreads()
ArrayList<Point>	getPeaks() If you want to get subpixel-localized peaks, call getSubpixelPeaks() directly.
ArrayList<RefinedPeak<Point>>	getSubpixelPeaks()
DogDetection.TypedDogDetection<?>	getTypedDogDetection()
void	setExecutorService(ExecutorService service)
void	setImageSigma(double imageSigma)
void	setKeepDoGImg(boolean keepDoGImg)
void	setMinf(double minf)
void	setMinPeakValue(double minPeakValue)
void	setNormalizeMinPeakValue(boolean normalizeMinPeakValue)
void	setNumThreads(int numThreads)

Methods inherited from class java.lang.Object

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

Field Detail

input

protected final RandomAccessible<T extends RealType<T> & NativeType<T>> input

interval

protected final Interval interval

sigmaSmaller

protected final double sigmaSmaller

sigmaLarger

protected final double sigmaLarger

pixelSize

protected final double[] pixelSize

typedDogDetection

protected final DogDetection.TypedDogDetection<?> typedDogDetection

imageSigma

protected double imageSigma

minf

protected double minf

extremaType

protected DogDetection.ExtremaType extremaType

minPeakValue

protected double minPeakValue

normalizeMinPeakValue

protected boolean normalizeMinPeakValue

keepDoGImg

protected boolean keepDoGImg

numThreads

protected int numThreads

Constructor Detail

DogDetection

public DogDetection(RandomAccessibleInterval<T> input,
                    double[] calibration,
                    double sigmaSmaller,
                    double sigmaLarger,
                    DogDetection.ExtremaType extremaType,
                    double minPeakValue,
                    boolean normalizeMinPeakValue)

DogDetection

public DogDetection(RandomAccessible<T> input,
                    Interval interval,
                    double[] calibration,
                    double sigmaSmaller,
                    double sigmaLarger,
                    DogDetection.ExtremaType extremaType,
                    double minPeakValue,
                    boolean normalizeMinPeakValue)

Sets up a DogDetection with the specified parameters (does not do any computation yet). If the input image is of type DoubleType, DoubleType will be used for computing the Difference-of-Gaussian. In all other cases, FloatType will be used).

Parameters:

input - the input image.

interval - which interval of the input image to process

calibration - The calibration, i.e., the voxel sizes in some unit for the input image.

sigmaSmaller - sigma for the smaller scale in the same units as calibration.

sigmaLarger - sigma for the larger scale in the same units as calibration.

extremaType - which type of extrema (minima, maxima) to detect. Note that minima in the Difference-of-Gaussian correspond to bright blobs on dark background. Maxima correspond to dark blobs on bright background.

minPeakValue - threshold value for detected extrema. Maxima below minPeakValue or minima above -minPeakValue will be disregarded.

normalizeMinPeakValue - Whether the peak value should be normalized. The Difference-of-Gaussian is an approximation of the scale-normalized Laplacian-of-Gaussian, with a factor of f = sigmaSmaller / (sigmaLarger - sigmaSmaller). If normalizeMinPeakValue=true, the minPeakValue will be divided by f (which is equivalent to scaling the DoG by f).

DogDetection

public DogDetection(RandomAccessible<T> input,
                    Interval interval,
                    double[] calibration,
                    double sigmaSmaller,
                    double sigmaLarger,
                    DogDetection.ExtremaType extremaType,
                    double minPeakValue,
                    boolean normalizeMinPeakValue,
                    F computationType)

Sets up a DogDetection with the specified parameters (does not do any computation yet).

Parameters:

input - the input image.

interval - which interval of the input image to process

calibration - The calibration, i.e., the voxel sizes in some unit for the input image.

sigmaSmaller - sigma for the smaller scale in the same units as calibration.

sigmaLarger - sigma for the larger scale in the same units as calibration.

extremaType - which type of extrema (minima, maxima) to detect. Note that minima in the Difference-of-Gaussian correspond to bright blobs on dark background. Maxima correspond to dark blobs on bright background.

minPeakValue - threshold value for detected extrema. Maxima below minPeakValue or minima above -minPeakValue will be disregarded.

normalizeMinPeakValue - Whether the peak value should be normalized. The Difference-of-Gaussian is an approximation of the scale-normalized Laplacian-of-Gaussian, with a factor of f = sigmaSmaller / (sigmaLarger - sigmaSmaller). If normalizeMinPeakValue=true, the minPeakValue will be divided by f (which is equivalent to scaling the DoG by f).

computationType - The type to use for computing the Difference-of-Gaussian.

Method Detail

getPeaks

public ArrayList<Point> getPeaks()

If you want to get subpixel-localized peaks, call getSubpixelPeaks() directly.

getSubpixelPeaks

public ArrayList<RefinedPeak<Point>> getSubpixelPeaks()

setImageSigma

public void setImageSigma(double imageSigma)

setMinf

public void setMinf(double minf)

setMinPeakValue

public void setMinPeakValue(double minPeakValue)

setNormalizeMinPeakValue

public void setNormalizeMinPeakValue(boolean normalizeMinPeakValue)

setKeepDoGImg

public void setKeepDoGImg(boolean keepDoGImg)

setNumThreads

public void setNumThreads(int numThreads)

getImageSigma

public double getImageSigma()

getMinf

public double getMinf()

getMinPeakValue

public double getMinPeakValue()

getNormalizeMinPeakValue

public boolean getNormalizeMinPeakValue()

getKeepDoGImg

public boolean getKeepDoGImg()

getNumThreads

public int getNumThreads()

setExecutorService

public void setExecutorService(ExecutorService service)

getTypedDogDetection

public DogDetection.TypedDogDetection<?> getTypedDogDetection()