net.imglib2.realtransform

Interface RealTransform

All Known Subinterfaces:

AffineGet, DifferentiableRealTransform, InvertibleRealTransform, ScaleAndTranslationGet, ScaleGet, TranslationGet

All Known Implementing Classes:

AbstractAffineTransform, AbstractDifferentiableRealTransform, AbstractRealTransformSequence, AbstractScale, AbstractTranslation, AffineTransform, AffineTransform2D, AffineTransform3D, ClippedTransitionRealTransform, DeformationFieldTransform, DisplacementFieldTransform, ExplicitInvertibleRealTransform, HomographyTransform2D, InterpolatedRealTransform, InversePerspective3D, InverseRealTransform, InverseRealTransformGradientDescent, InvertibleDeformationFieldTransform, InvertibleDisplacementFieldTransform, InvertibleRealTransformSequence, InvertibleTransformByGradientDescent, Perspective3D, PolarToCartesianTransform2D, PolynomialTransform2D, PositionFieldTransform, RealTransformFiniteDerivatives, RealTransformSequence, RegularizedDifferentiableRealTransform, Scale, Scale2D, Scale3D, ScaleAndTranslation, ScaledPolarToTranslatedCartesianTransform2D, SphericalToCartesianTransform3D, ThinplateSplineTransform, Translation, Translation2D, Translation3D, WrappedIterativeInvertibleRealTransform

public interface RealTransform

Transformation from Rⁿ to R^m.

Applying the transformation to an n-dimensional source vector yields an m-dimensional target vector.

Source and target vectors are required to be at least n- or m-dimensional respectively. The RealTransform is expected to leave all dimensions beyond n-1 in the source vector and m -1 in the target vector unchanged.

Implementations must support an in-place apply (i.e., passing the same object as both source and target).

Author:

Tobias Pietzsch, Stephan Saalfeld

Method Summary

Modifier and Type	Method and Description
void	apply(double[] source, double[] target) Apply the RealTransform to a source vector to obtain a target vector.
default void	apply(float[] source, float[] target) Deprecated. use double precision instead
default void	apply(long[] source, double[] target) Apply the RealTransform to a source vector to obtain a target vector.
void	apply(net.imglib2.RealLocalizable source, net.imglib2.RealPositionable target) Apply the RealTransform to a source RealLocalizable to obtain a target RealPositionable.
default double	applyDimension(double[] source, int d) Apply the RealTransform to a source vector to obtain one dimension of the resulting target vector.
default double	applyDimension(long[] source, int d) Apply the RealTransform to a source vector to obtain one dimension of the resulting target vector.
default double	applyDimension(net.imglib2.RealLocalizable source, int d) Apply the RealTransform to a source vector to obtain one dimension of the resulting target vector.
default net.imglib2.RealInterval	boundingInterval(net.imglib2.RealInterval interval, IntervalSamplingMethod samplingMethod) Estimate the RealInterval that bounds the given RealInterval after being transformed by a RealTransform.
RealTransform	copy() Create a deep copy of this RealTransform.
default boolean	isIdentity() Check if the RealTransform is identity.
int	numSourceDimensions() Returns n, the minimal number of dimension of the source vector.
int	numTargetDimensions() Returns m, the minimal dimension of the target vector.

Method Detail

numSourceDimensions

int numSourceDimensions()

Returns n, the minimal number of dimension of the source vector.

Returns:

the dimension of the source vector.

numTargetDimensions

int numTargetDimensions()

Returns m, the minimal dimension of the target vector.

Returns:

the dimension of the target vector.

apply

void apply(double[] source,
           double[] target)

Apply the RealTransform to a source vector to obtain a target vector.

Parameters:

source - source coordinates, length must be >= numSourceDimensions()

target - set this to the target coordinates, length must be >= numTargetDimensions()

apply

default void apply(long[] source,
                   double[] target)

Apply the RealTransform to a source vector to obtain a target vector.

Parameters:

source - source coordinates, length must be >= numSourceDimensions()

target - set this to the target coordinates, length must be >= numTargetDimensions() TODO override for this to be more efficient

apply

@Deprecated
default void apply(float[] source,
                               float[] target)

Deprecated. use double precision instead

Apply the RealTransform to a source vector to obtain a target vector.

Parameters:

source - source coordinates, length must be >= numSourceDimensions()

target - set this to the target coordinates, length must be >= numTargetDimensions()

apply

void apply(net.imglib2.RealLocalizable source,
           net.imglib2.RealPositionable target)

Apply the RealTransform to a source RealLocalizable to obtain a target RealPositionable.

Parameters:

source - source coordinates, numDimensions() must be >= numSourceDimensions()

target - set this to the target coordinates, numDimensions() must be >= numTargetDimensions()

applyDimension

default double applyDimension(double[] source,
                              int d)

Apply the RealTransform to a source vector to obtain one dimension of the resulting target vector.

Parameters:

source - source coordinates, length must be >= numSourceDimensions()

d - target dimension, must be >= numTargetDimensions()

Returns:

the d-th dimension of the transformed target vector TODO This default implementation creates and calculates the entire target vector and extracts the dimension from that vector. Override for this to be more efficient.

applyDimension

default double applyDimension(long[] source,
                              int d)

Apply the RealTransform to a source vector to obtain one dimension of the resulting target vector.

Parameters:

source - source coordinates, length must be >= numSourceDimensions()

d - target dimension, must be >= numTargetDimensions()

Returns:

the d-th dimension of the transformed target vector TODO This default implementation creates and calculates the entire target vector and extracts the dimension from that vector. Override for this to be more efficient.

applyDimension

default double applyDimension(net.imglib2.RealLocalizable source,
                              int d)

Apply the RealTransform to a source vector to obtain one dimension of the resulting target vector.

Parameters:

source - source coordinates, EuclideanSpace.numDimensions() must be >= numSourceDimensions()

d - target dimension, must be >= numTargetDimensions()

Returns:

the d-th dimension of the transformed target vector TODO This default implementation creates and calculates the entire target vector and extracts the dimension from that vector. Override for this to be more efficient.

copy

RealTransform copy()

Create a deep copy of this RealTransform.

Deep copying is required to make sure that stateful RealTransforms can be duplicated for concurrent code.

Returns:

deep copy

isIdentity

default boolean isIdentity()

Check if the RealTransform is identity.

Returns:

true if identity

boundingInterval

default net.imglib2.RealInterval boundingInterval(net.imglib2.RealInterval interval,
                                                  IntervalSamplingMethod samplingMethod)

Estimate the RealInterval that bounds the given RealInterval after being transformed by a RealTransform.

For arbitrary transformations, it is not necessarily possible to directly calculate the resulting bounding box trivially, therefore, a sampling methods for coordinates in the source interval must be provided. RealTransforms that can calculate the bounding interval more efficiently than by sampling coordinates are encouraged to override this method and to ignore the provided sampling method.

Parameters:

interval - the real interval

samplingMethod - the method used to sample coordinates of the source interval

Returns:

the bounding interval