Package qupath.opencv.tools

Class MultiscaleFeatures.MultiscaleResultsBuilder

java.lang.Object

qupath.opencv.tools.MultiscaleFeatures.MultiscaleResultsBuilder

Enclosing class:

MultiscaleFeatures

public static class MultiscaleFeatures.MultiscaleResultsBuilder
extends Object

Helper-class for computing pixel-features at a specified scale.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	MultiscaleFeatures.MultiscaleResultsBuilder.FeatureMap	Helper map implementation that provides access to MultiscaleFeatures.Hessian if needed.

Constructor Summary

Constructors

Constructor	Description
MultiscaleResultsBuilder()	Default constructor.
MultiscaleResultsBuilder(Collection<MultiscaleFeatures.MultiscaleFeature> features)	Constructor prepared to calculate specified features.

Method Summary

Modifier and Type	Method	Description
List<MultiscaleFeatures.MultiscaleResultsBuilder.FeatureMap>	build(List<Mat> mats)	Calculate results as a list of maps connecting features and Mats for all slices of a z-stack.
MultiscaleFeatures.MultiscaleResultsBuilder.FeatureMap	build(List<Mat> mats, int ind)	Calculate results for a map of features and Mats for one slice of a z-stack.
MultiscaleFeatures.MultiscaleResultsBuilder.FeatureMap	build(Mat mat)	Calculate results for a single Mat.
MultiscaleFeatures.MultiscaleResultsBuilder	gaussianSmoothed(boolean calculate)	Calculate the Gaussian-smoothed image.
MultiscaleFeatures.MultiscaleResultsBuilder	gradientMagnitude(boolean calculate)	Calculate the gradient magnitude.
MultiscaleFeatures.MultiscaleResultsBuilder	hessianDeterminant(boolean calculate)	Calculate the determinant of the Hessian matrix per pixel.
MultiscaleFeatures.MultiscaleResultsBuilder	hessianEigenvalues(boolean calculate)	Calculate the eigenvalues of the Hessian matrix per pixel.
MultiscaleFeatures.MultiscaleResultsBuilder	laplacianOfGaussian(boolean calculate)	Calculate the Laplacian of Gaussian image.
MultiscaleFeatures.MultiscaleResultsBuilder	paddingXY(int padding)	Specify the number of pixels that the input image is padded (left, right, above, below).
MultiscaleFeatures.MultiscaleResultsBuilder	pixelCalibration(PixelCalibration cal, double downsampleXY)	Set the pixel calibration, with optional x,y scaling.
MultiscaleFeatures.MultiscaleResultsBuilder	retainHessian(boolean retain)	Optionally retain the Hessian matrix per pixel.
MultiscaleFeatures.MultiscaleResultsBuilder	sigmaX(double sigma)	Set all Gaussian sigma values for the horizontal filter.
MultiscaleFeatures.MultiscaleResultsBuilder	sigmaXY(double sigma)	Set all Gaussian sigma values (x, y) to the same value.
MultiscaleFeatures.MultiscaleResultsBuilder	sigmaY(double sigma)	Set all Gaussian sigma values for the vertical filter.
MultiscaleFeatures.MultiscaleResultsBuilder	sigmaZ(double sigma)	Set all Gaussian sigma values for the z-dimension filter.
MultiscaleFeatures.MultiscaleResultsBuilder	structureTensorEigenvalues(boolean calculate)	Calculate the eigenvalues of the structure tensor (not yet implemented!).
MultiscaleFeatures.MultiscaleResultsBuilder	weightedStdDev(boolean calculate)	Calculate a Gaussian-weighted standard deviation.

Methods inherited from class java.lang.Object

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

Constructor Details

MultiscaleResultsBuilder

public MultiscaleResultsBuilder()

Default constructor.

MultiscaleResultsBuilder

public MultiscaleResultsBuilder(Collection<MultiscaleFeatures.MultiscaleFeature> features)

Constructor prepared to calculate specified features.

Parameters:

features -

Method Details

gaussianSmoothed

public MultiscaleFeatures.MultiscaleResultsBuilder gaussianSmoothed(boolean calculate)

Calculate the Gaussian-smoothed image.

Parameters:

calculate -

Returns:

paddingXY

public MultiscaleFeatures.MultiscaleResultsBuilder paddingXY(int padding)

Specify the number of pixels that the input image is padded (left, right, above, below). This padding will be stripped prior to outputting the results. Default is 0.

Parameters:

padding -

Returns:

weightedStdDev

public MultiscaleFeatures.MultiscaleResultsBuilder weightedStdDev(boolean calculate)

Calculate a Gaussian-weighted standard deviation.

Parameters:

calculate -

Returns:

gradientMagnitude

public MultiscaleFeatures.MultiscaleResultsBuilder gradientMagnitude(boolean calculate)

Calculate the gradient magnitude.

Parameters:

calculate -

Returns:

structureTensorEigenvalues

public MultiscaleFeatures.MultiscaleResultsBuilder structureTensorEigenvalues(boolean calculate)

Calculate the eigenvalues of the structure tensor (not yet implemented!).

Parameters:

calculate -

Returns:

laplacianOfGaussian

public MultiscaleFeatures.MultiscaleResultsBuilder laplacianOfGaussian(boolean calculate)

Calculate the Laplacian of Gaussian image.

Parameters:

calculate -

Returns:

hessianEigenvalues

public MultiscaleFeatures.MultiscaleResultsBuilder hessianEigenvalues(boolean calculate)

Calculate the eigenvalues of the Hessian matrix per pixel.

Parameters:

calculate -

Returns:

retainHessian

public MultiscaleFeatures.MultiscaleResultsBuilder retainHessian(boolean retain)

Optionally retain the Hessian matrix per pixel. This provides an opportunity to request eigenvectors, and/or to sort eigenvalues by absolute value.

Parameters:

retain -

Returns:

hessianDeterminant

public MultiscaleFeatures.MultiscaleResultsBuilder hessianDeterminant(boolean calculate)

Calculate the determinant of the Hessian matrix per pixel.

Parameters:

calculate -

Returns:

pixelCalibration

public MultiscaleFeatures.MultiscaleResultsBuilder pixelCalibration(PixelCalibration cal,
 double downsampleXY)

Set the pixel calibration, with optional x,y scaling.

If available, this will convert the units in which Gaussian sigma values are applied. This helps apply an isotropic filtering more easily, by specifying a single sigma value and supplying the pixel calibration so that any differences in pixel dimensions is automatically adjusted for.

DownsampleXY may be further used to scale the pixel units, meaning that even when working with a downsampled image it is possible to pass the original PixelCalibration object and specify the level of downsampling (rather than need to create a new scaled PixelCalibration).

Parameters:

cal -

downsampleXY -

Returns:

See Also:

• sigmaXY(double)
• sigmaX(double)
• sigmaY(double)
• sigmaZ(double)

sigmaXY

public MultiscaleFeatures.MultiscaleResultsBuilder sigmaXY(double sigma)

Set all Gaussian sigma values (x, y) to the same value.

Note that this value is in pixels by default, or may be microns is supported by setting the pixel calibration.

Parameters:

sigma -

Returns:

See Also:

• pixelCalibration(PixelCalibration, double)
• sigmaX(double)
• sigmaY(double)

sigmaX

public MultiscaleFeatures.MultiscaleResultsBuilder sigmaX(double sigma)

Set all Gaussian sigma values for the horizontal filter.

Note that this value is in pixels by default, or may be microns is supported by setting the pixel calibration.

Parameters:

sigma -

Returns:

See Also:

• pixelCalibration(PixelCalibration, double)
• sigmaXY(double)
• sigmaY(double)
• sigmaZ(double)

sigmaY

public MultiscaleFeatures.MultiscaleResultsBuilder sigmaY(double sigma)

Set all Gaussian sigma values for the vertical filter.

Note that this value is in pixels by default, or may be microns is supported by setting the pixel calibration.

Parameters:

sigma -

Returns:

See Also:

• pixelCalibration(PixelCalibration, double)
• sigmaXY(double)
• sigmaX(double)
• sigmaZ(double)

sigmaZ

public MultiscaleFeatures.MultiscaleResultsBuilder sigmaZ(double sigma)

Set all Gaussian sigma values for the z-dimension filter.

Note that this value is in pixels by default, or may be microns is supported by setting the pixel calibration.

Parameters:

sigma -

Returns:

See Also:

• pixelCalibration(PixelCalibration, double)
• sigmaXY(double)
• sigmaX(double)
• sigmaY(double)

build

public MultiscaleFeatures.MultiscaleResultsBuilder.FeatureMap build(Mat mat)

Calculate results for a single Mat.

Parameters:

mat -

Returns:

build

public MultiscaleFeatures.MultiscaleResultsBuilder.FeatureMap build(List<Mat> mats,
 int ind)

Calculate results for a map of features and Mats for one slice of a z-stack.

Parameters:

mats - a list of mats, one for each slice of the z-stack

ind - the index of the slice to use

Returns:

build

public List<MultiscaleFeatures.MultiscaleResultsBuilder.FeatureMap> build(List<Mat> mats)

Calculate results as a list of maps connecting features and Mats for all slices of a z-stack.

Parameters:

mats -

Returns: