- * This method differs from - * {@link #fill(RandomAccessible, RandomAccessible, Localizable, Type, Shape)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.floodFill', type=Computer - * @param source - * input - * @param seed - * Start flood fill at this location. - * @param fillLabel - * Immutable. Value to be written into valid flood fill - * locations. - * @param shape - * Defines neighborhood that is considered for connected - * components, e.g. - * {@link net.imglib2.algorithm.neighborhood.DiamondShape} - * @param target - * {@link RandomAccessible} to be written into. May be the same - * as input. - * @param- * This method differs from - * {@link #fill(RandomAccessible, RandomAccessible, Localizable, Type, Shape, BiPredicate)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.floodFill', type=Computer - * @param source - * input - * @param seed - * Start flood fill at this location. - * @param fillLabel - * Immutable. Value to be written into valid flood fill - * locations. - * @param shape - * Defines neighborhood that is considered for connected - * components, e.g. - * {@link net.imglib2.algorithm.neighborhood.DiamondShape} - * @param filter - * Returns true if pixel has not been visited yet and should be - * written into. Returns false if target pixel has been visited - * or source pixel is not part of the same connected component. - * @param target - * {@link RandomAccessible} to be written into. May be the same - * as input. - * @param- * This method differs from - * {@link #fill(RandomAccessible, RandomAccessible, Localizable, Shape, BiPredicate, Consumer)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.floodFill', type=Computer - * @param source - * input - * @param seed - * Start flood fill at this location. - * @param shape - * Defines neighborhood that is considered for connected - * components, e.g. - * {@link net.imglib2.algorithm.neighborhood.DiamondShape} - * @param filter - * Returns true if pixel has not been visited yet and should be - * written into. Returns false if target pixel has been visited - * or source pixel is not part of the same connected component. - * @param writer - * Defines how fill label is written into target at current - * location. - * @param target - * {@link RandomAccessible} to be written into. May be the same - * as input. - * @param- * This method differs from - * {@link #gradientCentralDifference2(RandomAccessible, RandomAccessibleInterval, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='filter.gradientCentralDifference2', type=Computer - * @param source - * source image, has to provide valid data in the interval of the - * gradient image plus a one pixel border in dimension. - * @param dimension - * along which dimension the partial derivatives are computed - * @param gradient - * output image - */ - public static < T extends NumericType< T > > void gradientCentralDifference2( final RandomAccessible< T > source, final int dimension, final RandomAccessibleInterval< T > gradient) - { - gradientCentralDifference2( source, gradient, dimension ); - } - // parallel version... /** * Compute the partial derivative (central difference approximation) of source @@ -133,6 +107,7 @@ public static < T extends NumericType< T > > void gradientCentralDifference2( fi * @param es * {@link ExecutorService} providing workers for gradient * computation. Service is managed (created, shutdown) by caller. + * @implNote op name='filter.partialDerivative, filter.gradientCentralDifference', type=Computer2, priority='-1000.' */ public static < T extends NumericType< T > > void gradientCentralDifferenceParallel( final RandomAccessible< T > source, @@ -189,42 +164,6 @@ public static < T extends NumericType< T > > void gradientCentralDifferenceParal f.get(); } - /** - * Compute the partial derivative (central difference approximation) of source - * in a particular dimension: - * {@code d_f( x ) = ( f( x + e ) - f( x - e ) ) / 2}, - * where {@code e} is the unit vector along that dimension. - *- * This method differs from - * {@link #gradientCentralDifferenceParallel(RandomAccessible, RandomAccessibleInterval, int, int, ExecutorService)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='filter.gradientCentralDifferenceParallel', type=Computer - * @param source - * source image, has to provide valid data in the interval of the - * gradient image plus a one pixel border in dimension. - * @param dimension - * along which dimension the partial derivatives are computed - * @param nTasks - * Number of tasks for gradient computation. - * @param es - * {@link ExecutorService} providing workers for gradient - * computation. Service is managed (created, shutdown) by caller. - * @param gradient - * output image - */ - public static < T extends NumericType< T > > void gradientCentralDifferenceParallel( - final RandomAccessible< T > source, - final int dimension, - final int nTasks, - final ExecutorService es, - final RandomAccessibleInterval< T > gradient ) throws InterruptedException, ExecutionException - { - gradientCentralDifferenceParallel( source, gradient, dimension, nTasks, es ); - } - // fast version /** * Compute the partial derivative (central difference approximation) of source @@ -239,6 +178,7 @@ public static < T extends NumericType< T > > void gradientCentralDifferenceParal * output image * @param dimension * along which dimension the partial derivatives are computed + * @implNote op name='filter.partialDerivative, filter.gradientCentralDifference', type=Computer2, priority='-1000.' */ public static < T extends NumericType< T > > void gradientCentralDifference( final RandomAccessible< T > source, final RandomAccessibleInterval< T > result, final int dimension ) @@ -253,33 +193,6 @@ public static < T extends NumericType< T > > void gradientCentralDifference( fin } ); } - /** - * Compute the partial derivative (central difference approximation) of source - * in a particular dimension: - * {@code d_f( x ) = ( f( x + e ) - f( x - e ) ) / 2}, - * where {@code e} is the unit vector along that dimension. - *- * This method differs from - * {@link #gradientCentralDifference(RandomAccessible, RandomAccessibleInterval, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='filter.gradientCentralDifference', type=Computer - * @param source - * source image, has to provide valid data in the interval of the - * gradient image plus a one pixel border in dimension. - * @param dimension - * along which dimension the partial derivatives are computed - * @param result - * output image - */ - public static < T extends NumericType< T > > void gradientCentralDifference( final RandomAccessible< T > source, - final int dimension, final RandomAccessibleInterval< T > result) - { - gradientCentralDifference( source, result, dimension ); - } - /** * Compute the backward difference of source in a particular dimension: * {@code d_f( x ) = ( f( x ) - f( x - e ) )} @@ -289,6 +202,7 @@ public static < T extends NumericType< T > > void gradientCentralDifference( fin * the gradient image plus a one pixel border in dimension. * @param result output image * @param dimension along which dimension the partial derivatives are computed + * @implNote op name='filter.gradientBackwardDifference', type=Computer2 */ public static < T extends NumericType< T > > void gradientBackwardDifference( final RandomAccessible< T > source, final RandomAccessibleInterval< T > result, final int dimension ) @@ -302,29 +216,6 @@ public static < T extends NumericType< T > > void gradientBackwardDifference( fi } ); } - /** - * Compute the backward difference of source in a particular dimension: - * {@code d_f( x ) = ( f( x ) - f( x - e ) )} - * where {@code e} is the unit vector along that dimension - *- * This method differs from - * {@link #gradientBackwardDifference(RandomAccessible, RandomAccessibleInterval, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='filter.gradientBackwardDifference', type=Computer - * @param source source image, has to provide valid data in the interval of - * the gradient image plus a one pixel border in dimension. - * @param dimension along which dimension the partial derivatives are computed - * @param result output image - */ - public static < T extends NumericType< T > > void gradientBackwardDifference( final RandomAccessible< T > source, - final int dimension, final RandomAccessibleInterval< T > result) - { - gradientBackwardDifference( source, result, dimension ); - } - /** * Compute the forward difference of source in a particular dimension: * {@code d_f( x ) = ( f( x + e ) - f( x ) )} @@ -334,6 +225,7 @@ public static < T extends NumericType< T > > void gradientBackwardDifference( fi * the gradient image plus a one pixel border in dimension. * @param result output image * @param dimension along which dimension the partial derivatives are computed + * @implNote op name='filter.gradientForwardDifference', type=Computer2 */ public static < T extends NumericType< T > > void gradientForwardDifference( final RandomAccessible< T > source, final RandomAccessibleInterval< T > result, final int dimension ) @@ -347,26 +239,4 @@ public static < T extends NumericType< T > > void gradientForwardDifference( fin } ); } - /** - * Compute the forward difference of source in a particular dimension: - * {@code d_f( x ) = ( f( x + e ) - f( x ) )} - * where {@code e} is the unit vector along that dimension - *- * This method differs from - * {@link #gradientForwardDifference(RandomAccessible, RandomAccessibleInterval, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='filter.gradientForwardDifference', type=Computer - * @param source source image, has to provide valid data in the interval of - * the gradient image plus a one pixel border in dimension. - * @param dimension along which dimension the partial derivatives are computed - * @param result output image - */ - public static < T extends NumericType< T > > void gradientForwardDifference( final RandomAccessible< T > source, - final int dimension, final RandomAccessibleInterval< T > result) - { - gradientForwardDifference( source, result, dimension ); - } } diff --git a/src/main/java/net/imglib2/algorithm/hough/HoughTransforms.java b/src/main/java/net/imglib2/algorithm/hough/HoughTransforms.java index 1557a7a2b..7d73d157f 100644 --- a/src/main/java/net/imglib2/algorithm/hough/HoughTransforms.java +++ b/src/main/java/net/imglib2/algorithm/hough/HoughTransforms.java @@ -106,7 +106,7 @@ private static int defaultRho( final Dimensions dimensions ) * Returns the size of the vote space output image given an input * {@link RandomAccessibleInterval}. * - * @implNote op names='filter.hough.getVotespaceSize', type=Function + * @implNote op names='features.hough.getVotespaceSize', type=Function * @param dimensions * - the {@link Dimensions} over which the Hough Line Transform * will be run @@ -121,7 +121,7 @@ public static long[] getVotespaceSize( final Dimensions dimensions ) * Returns the size of the vote space output image given an input * {@link RandomAccessibleInterval}. * - * @implNote op names='filter.hough.getVotespaceSize', type=Function + * @implNote op names='features.hough.getVotespaceSize', type=Function * @param dimensions * - the {@link Dimensions} over which the Hough Line Transform * will be run @@ -138,7 +138,7 @@ public static long[] getVotespaceSize( final Dimensions dimensions, final int nT * Returns the size of the voteSpace output image given desired {@code nRho} * and {@code nTheta} values. * - * @implNote op names='filter.hough.getVotespaceSize', type=Function + * @implNote op names='features.hough.getVotespaceSize', type=Function * @param nRho * - the number of bins for rho resolution * @param nTheta @@ -153,7 +153,7 @@ public static long[] getVotespaceSize( final int nRho, final int nTheta ) /** * Pick vote space peaks with a {@link LocalExtrema}. * - * @implNote op names='filter.hough.pickLinePeaks', type=Function + * @implNote op names='features.hough.pickPeaks', type=Function * @param voteSpace * - the {@link RandomAccessibleInterval} containing the output * of a Hough Transform vote @@ -175,7 +175,7 @@ public static < T extends IntegerType< T > > List< Point > pickLinePeaks( /** * Pick vote space peaks with a {@link LocalExtrema}. * - * @implNote op names='filter.hough.pickLinePeaks', type=Function + * @implNote op names='features.hough.pickPeaks', type=Function * @param voteSpace * - the {@link RandomAccessibleInterval} containing the output * of a Hough Transform vote @@ -207,7 +207,7 @@ public static < T extends Comparable< T > > List< Point > pickLinePeaks( * Runs a Hough Line Tranform on an image and populates the vote space * parameter with the results. * - * @implNote op names='filter.hough.voteLines', + * @implNote op names='features.hough.voteLines', * type=Computer * @param input * - the {@link RandomAccessibleInterval} to run the Hough Line @@ -235,6 +235,7 @@ public static < T extends Comparable< T >, U extends IntegerType< U > > void vot * are stored * @param nTheta * - the number of bins for theta resolution + * @implNote op name='features.hough.voteLines', type=Computer2 */ public static < T extends Comparable< T >, U extends IntegerType< U > > void voteLines( final RandomAccessibleInterval< T > input, @@ -265,34 +266,6 @@ private static- * This method differs from - * {@link #voteLines(RandomAccessibleInterval, RandomAccessibleInterval, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='filter.hough.voteLines', type=Computer - * @param input - * - the {@link RandomAccessibleInterval} to run the Hough Line - * Transform over - * @param nTheta - * - the number of bins for theta resolution - * @param votespace - * - the {@link RandomAccessibleInterval} in which the results - * are stored - */ - public static < T extends Comparable< T >, U extends IntegerType< U > > void voteLines( - final RandomAccessibleInterval< T > input, - final int nTheta, - final RandomAccessibleInterval< U > votespace) - { - voteLines( input, votespace, nTheta); - } - /** * Runs a Hough Line Tranform on an image and populates the vote space * parameter with the results. @@ -307,6 +280,7 @@ public static < T extends Comparable< T >, U extends IntegerType< U > > void vot * - the number of bins for theta resolution * @param nRho * - the number of bins for rho resolution + * @implNote op names='features.hough.voteLines', type=Computer2 */ public static < T extends Comparable< T >, U extends IntegerType< U > > void voteLines( final RandomAccessibleInterval< T > input, @@ -317,37 +291,6 @@ public static < T extends Comparable< T >, U extends IntegerType< U > > void vot voteLines( input, votespace, nTheta, nRho, getTypeFromInterval( input ) ); } - /** - * Runs a Hough Line Tranform on an image and populates the vote space - * parameter with the results. - *- * This method differs from - * {@link #voteLines(RandomAccessibleInterval, RandomAccessibleInterval, int, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op names='filter.hough.voteLines', type=Computer - * @param input - * - the {@link RandomAccessibleInterval} to run the Hough Line - * Transform over - * @param votespace - * - the {@link RandomAccessibleInterval} in which the results - * are stored - * @param nTheta - * - the number of bins for theta resolution - * @param nRho - * - the number of bins for rho resolution - */ - public static < T extends Comparable< T >, U extends IntegerType< U > > void voteLines( - final RandomAccessibleInterval< T > input, - final int nTheta, - final int nRho, - final RandomAccessibleInterval< U > votespace) - { - voteLines(input, votespace, nTheta, nRho); - } - /** * Runs a Hough Line Tranform on an image and populates the vote space * parameter with the results. @@ -365,6 +308,7 @@ public static < T extends Comparable< T >, U extends IntegerType< U > > void vot * @param threshold * - the minimum value allowed by the populator. Any input less * than this value will be disregarded by the populator. + * @implNote op name='features.hough.voteLines', type=Computer2 */ public static < T extends Comparable< T >, U extends IntegerType< U > > void voteLines( final RandomAccessibleInterval< T > input, @@ -378,42 +322,6 @@ public static < T extends Comparable< T >, U extends IntegerType< U > > void vot voteLines( input, votespace, nTheta, nRho, p ); } - /** - * Runs a Hough Line Tranform on an image and populates the vote space - * parameter with the results. - *- * This method differs from - * {@link #voteLines(RandomAccessibleInterval, RandomAccessibleInterval, int, int, Comparable)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='filter.hough.voteLines', type=Computer - * @param input - * - the {@link RandomAccessibleInterval} to run the Hough Line - * Transform over - * @param nTheta - * - the number of bins for theta resolution - * @param nRho - * - the number of bins for rho resolution - * @param threshold - * - the minimum value allowed by the populator. Any input less - * than this value will be disregarded by the populator. - * @param votespace - * - the {@link RandomAccessibleInterval} in which the results - * are stored - */ - public static < T extends Comparable< T >, U extends IntegerType< U > > void voteLines( - final RandomAccessibleInterval< T > input, - final int nTheta, - final int nRho, - final T threshold, - final RandomAccessibleInterval< U > votespace) - - { - voteLines(input, votespace, nTheta, nRho, threshold); - } - /** * * Runs a Hough Line Tranform on an image and populates the vote space @@ -468,6 +376,7 @@ public static < T extends Comparable< T >, U extends IntegerType< U > > void vot * above the minimum value allowed by the populator. Any input * less than or equal to this value will be disregarded by the * populator. + * @implNote op name='features.hough.voteLines', type=Computer2 */ public static < T, U extends IntegerType< U > > void voteLines( final RandomAccessibleInterval< T > input, @@ -527,85 +436,13 @@ public static < T, U extends IntegerType< U > > void voteLines( } } - /** - * - * Runs a Hough Line Tranform on an image and populates the vote space - * parameter with the results. - *- * Vote space here has two dimensions: {@code rho} and {@code theta}. - * {@code theta} is measured in radians {@code [-pi/2 pi/2)}, {@code rho} is - * measured in {@code [-rhoMax, rhoMax)}. - *
- *- * Lines are modeled as - *
- * - *- * l(t) = | x | = rho * | cos(theta) | + t * | sin(theta) | - * | y | | -sin(theta) | | cos(theta) | - *- *
- * In other words, {@code rho} represents the signed minimum distance from - * the image origin to the line, and {@code theta} indicates the angle - * between the row-axis and the minimum offset vector. - *
- *- * For a given point, then, votes are placed along the curve - *
- * - *- * rho = y * sin( theta ) + x * cos( theta ) - *- *
- * It is important to note that the interval of the first dimension of the - * vote space image is NOT {@code [-maxRho, maxRho)} but instead - * {@code [0, maxRho * 2)}; the same applies to the second dimension of the - * vote space as well. Thus if {@link HoughTransforms#pickLinePeaks} is not - * used to retrieve the maxima from the vote space, the vote space will have - * to be translated by {@code -maxRho} in dimension 0 to get the correct - * {@code rho} and {@code theta} values from the vote space. - *
- *- * This method differs from - * {@link #voteLines(RandomAccessibleInterval, RandomAccessibleInterval, int, int, Predicate)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='filter.hough.voteLines', type=Computer - * @param input - * - the {@link RandomAccessibleInterval} to run the Hough Line - * Transform over - * @param nTheta - * - the number of bins for theta resolution - * @param nRho - * - the number of bins for rho resolution - * @param filter - * - a {@link Predicate} judging whether or not the a value is - * above the minimum value allowed by the populator. Any input - * less than or equal to this value will be disregarded by the - * populator. - * @param votespace - * - the {@link RandomAccessibleInterval} in which the results - * are stored - */ - public static < T, U extends IntegerType< U > > void voteLines( - final RandomAccessibleInterval< T > input, - final int nTheta, - final int nRho, - final Predicate< T > filter, - final RandomAccessibleInterval< U > votespace) - { - voteLines(input, votespace, nTheta, nRho, filter); - } - /** * Method used to convert the {rho, theta} output of the * {@link HoughTransforms#voteLines} algorithm into a more useful * y-intercept value. Used with {@link HoughTransforms#getSlope} to create * line equations. * - * @implNote op name='filter.hough.getIntercept', type=Function + * @implNote op name='features.hough.intercept', type=Function * @param rho * - the {@code rho} of the line * @param theta @@ -624,7 +461,7 @@ public static double getIntercept( final long rho, final long theta ) * value. Used with {@link HoughTransforms#getIntercept} to create line * equations. * - * @implNote op name='filter.hough.getSlope', type=Function + * @implNote op name='features.hough.slope', type=Function * @param theta * - the {@code theta} of the line * @return {@code double} - the y-intercept of the line diff --git a/src/main/java/net/imglib2/algorithm/labeling/ConnectedComponentAnalysis.java b/src/main/java/net/imglib2/algorithm/labeling/ConnectedComponentAnalysis.java index 7d76fbe32..0a08f7bd6 100644 --- a/src/main/java/net/imglib2/algorithm/labeling/ConnectedComponentAnalysis.java +++ b/src/main/java/net/imglib2/algorithm/labeling/ConnectedComponentAnalysis.java @@ -117,7 +117,7 @@ public static < T > ToLongBiFunction< Localizable, T > idFromIntervalIndexer( fi * generalization for higher dimenions over a binary mask. {@code mask} and * {@code labeling} are expected to have equal min and max. * - * @implNote op name='labeling.connectedComponents', type=Computer + * @implNote op name='labeling.cca', type=Computer * @param mask * Boolean mask to distinguish foreground ({@code true}) from * background ({@code false}). @@ -155,6 +155,7 @@ public static < B extends BooleanType< B >, L extends IntegerType< L > > void co * ({@link DiamondShape}), 8-neighborhood * ({@link RectangleNeighborhood}) and their generalisations for * higher dimensions. + * @implNote op name='labeling.cca', type=Computer2 */ public static < B extends BooleanType< B >, L extends IntegerType< L > > void connectedComponents( final RandomAccessibleInterval< B > mask, @@ -171,44 +172,6 @@ public static < B extends BooleanType< B >, L extends IntegerType< L > > void co new StartAtOneIdForNextSet() ); } - /** - * - * Implementation of connected component analysis that uses - * {@link IntArrayRankedUnionFind} to find sets of pixels that are connected - * with respect to a neighborhood ({@code shape}) over a binary mask. - * {@code mask} - * and {@code labeling} are expected to have equal min and max. - *- * This method differs from - * {@link #connectedComponents(RandomAccessibleInterval, RandomAccessibleInterval, Shape)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='labeling.connectedComponents', type=Computer - * @param mask - * Boolean mask to distinguish foreground ({@code true}) from - * background ({@code false}). - * @param shape - * Connectivity of connected components, e.g. 4-neighborhood - * ({@link DiamondShape}), 8-neighborhood - * ({@link RectangleNeighborhood}) and their generalisations for - * higher dimensions. - * @param labeling - * Output parameter to store labeling: background pixels are - * labeled zero, foreground pixels are greater than zero: 1, 2, - * ..., N. Note that initially all pixels are expected to be zero - * as background values will not be written. - */ - public static < B extends BooleanType< B >, L extends IntegerType< L > > void connectedComponents( - final RandomAccessibleInterval< B > mask, - final Shape shape, - final RandomAccessibleInterval< L > labeling) - { - connectedComponents(mask, labeling, shape); - } - - /** * * Implementation of connected component analysis that uses @@ -239,6 +202,7 @@ public static < B extends BooleanType< B >, L extends IntegerType< L > > void co * @param idForSet * Create id for a set from the root id of a set. Multiple calls * with the same argument should always return the same result. + * @implNote op name='labeling.cca', type=Computer2 */ public static < B extends BooleanType< B >, L extends IntegerType< L > > void connectedComponents( final RandomAccessibleInterval< B > mask, @@ -255,55 +219,6 @@ public static < B extends BooleanType< B >, L extends IntegerType< L > > void co UnionFind.relabel( mask, labeling, uf, idForPixel, idForSet ); } - /** - * - * Implementation of connected component analysis that uses - * {@link UnionFind} to find sets of pixels that are connected with respect - * to a neighborhood ({@code shape}) over a binary mask. {@code mask} and - * {@code labeling} are expected to have equal min and max. - *- * This method differs from - * {@link #connectedComponents(RandomAccessibleInterval, RandomAccessibleInterval, Shape, LongFunction, ToLongBiFunction, LongUnaryOperator)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='labeling.connectedComponents', type=Computer - * @param mask - * Boolean mask to distinguish foreground ({@code true}) from - * background ({@code false}). - * @param shape - * Connectivity of connected components, e.g. 4-neighborhood - * ({@link DiamondShape}), 8-neighborhood - * ({@link RectangleNeighborhood}) and their generalisations for - * higher dimensions. - * @param unionFindFactory - * Creates appropriate {@link UnionFind} data structure for size - * of {@code labeling}, e.g. {@link IntArrayRankedUnionFind} of - * appropriate size. - * @param idForPixel - * Create id from pixel location and value. Multiple calls with - * the same argument should always return the same result. - * @param idForSet - * Create id for a set from the root id of a set. Multiple calls - * with the same argument should always return the same result. - * @param labeling - * Output parameter to store labeling: background pixels are - * labeled zero, foreground pixels are greater than zero: 1, 2, - * ..., N. Note that this is expected to be zero as background - * values will not be written. - */ - public static < B extends BooleanType< B >, L extends IntegerType< L > > void connectedComponents( - final RandomAccessibleInterval< B > mask, - final Shape shape, - final LongFunction< UnionFind > unionFindFactory, - final ToLongBiFunction< Localizable, L > idForPixel, - final LongUnaryOperator idForSet, - final RandomAccessibleInterval< L > labeling) - { - connectedComponents( mask, labeling, shape, unionFindFactory, idForPixel, idForSet ); - } - private static < B extends BooleanType< B >, L extends IntegerType< L > > UnionFind makeUnion( final RandomAccessibleInterval< B > mask, final RandomAccessibleInterval< L > labeling, diff --git a/src/main/java/net/imglib2/algorithm/labeling/ConnectedComponents.java b/src/main/java/net/imglib2/algorithm/labeling/ConnectedComponents.java index 531dc0b56..d6873f8b3 100644 --- a/src/main/java/net/imglib2/algorithm/labeling/ConnectedComponents.java +++ b/src/main/java/net/imglib2/algorithm/labeling/ConnectedComponents.java @@ -99,6 +99,7 @@ public CollectNeighborLabelsFactory getFactory() * @param se * structuring element to use. 8-connected or 4-connected * (respectively n-dimensional analog) + * @implNote op name='labeling.cca', type=Computer2 */ public static < T extends IntegerType< T >, L, I extends IntegerType< I > > void labelAllConnectedComponents( final RandomAccessible< T > input, @@ -112,41 +113,6 @@ public static < T extends IntegerType< T >, L, I extends IntegerType< I > > void service.shutdown(); } - /** - * Label all connected components in the given input image. In the output - * image, all background pixels will be labeled to {} and foreground - * components labeled as {1}, {2}, {3}, etc. where 1, 2, 3 are labels - * returned by {@code labelGenerator.next()}. {@code labelGenerator.next()} - * is called exactly n times if the input contains - * n connected components. - *- * This method differs from - * {@link #labelAllConnectedComponents(RandomAccessible, ImgLabeling, Iterator, StructuringElement)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='labeling.connectedComponents', type=Computer - * @param input - * input image with pixels != 0 belonging to foreground. - * @param labelGenerator - * produces labels for the connected components. - * @param se - * structuring element to use. 8-connected or 4-connected - * (respectively n-dimensional analog) - * @param labeling - * output labeling in which the connected components will be - * labeled. - */ - public static < T extends IntegerType< T >, L, I extends IntegerType< I > > void labelAllConnectedComponents( - final RandomAccessible< T > input, - final Iterator< L > labelGenerator, - final StructuringElement se, - final ImgLabeling< L, I > labeling) - { - labelAllConnectedComponents( input, labeling, labelGenerator, se ); - } - /** * Label all connected components in the given input image. In the output * image, all background pixels will be labeled to {} and foreground @@ -167,6 +133,7 @@ public static < T extends IntegerType< T >, L, I extends IntegerType< I > > void * (respectively n-dimensional analog) * @param service * service providing threads for multi-threading + * @implNote op name='labeling.cca', type=Computer2 */ public static < T extends IntegerType< T >, L, I extends IntegerType< I > > void labelAllConnectedComponents( final RandomAccessible< T > input, @@ -189,44 +156,6 @@ public static < T extends IntegerType< T >, L, I extends IntegerType< I > > void labeling.getMapping().setLabelSets( labelSets ); } - /** - * Label all connected components in the given input image. In the output - * image, all background pixels will be labeled to {} and foreground - * components labeled as {1}, {2}, {3}, etc. where 1, 2, 3 are labels - * returned by {@code labelGenerator.next()}. {@code labelGenerator.next()} - * is called exactly n times if the input contains - * n connected components. - *- * This method differs from - * {@link #labelAllConnectedComponents(RandomAccessible, ImgLabeling, Iterator, StructuringElement, ExecutorService)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='labeling.connectedComponents', type=Computer - * @param input - * input image with pixels != 0 belonging to foreground. - * @param labelGenerator - * produces labels for the connected components. - * @param se - * structuring element to use. 8-connected or 4-connected - * (respectively n-dimensional analog) - * @param service - * service providing threads for multi-threading - * @param labeling - * output labeling in which the connected components will be - * labeled. - */ - public static < T extends IntegerType< T >, L, I extends IntegerType< I > > void labelAllConnectedComponents( - final RandomAccessible< T > input, - final Iterator< L > labelGenerator, - final StructuringElement se, - final ExecutorService service, - final ImgLabeling< L, I > labeling) - { - labelAllConnectedComponents( input, labeling, labelGenerator, se, service ); - } - /** * "Label" all connected components in the given input image. In the output * image, all background pixels will be set to 0 and foreground components diff --git a/src/main/java/net/imglib2/algorithm/lazy/Lazy.java b/src/main/java/net/imglib2/algorithm/lazy/Lazy.java index ff64ed53b..eee487136 100644 --- a/src/main/java/net/imglib2/algorithm/lazy/Lazy.java +++ b/src/main/java/net/imglib2/algorithm/lazy/Lazy.java @@ -157,7 +157,7 @@ else if ( DoubleType.class.isInstance( type ) ) * Create a memory {@link CachedCellImg} with a cell generator * {@link Consumer}. * - * @implNote op name='create.cellImg, create', type=Function + * @implNote op name='create.img, engine.create', type=Function * @param targetInterval * @param blockSize * @param type diff --git a/src/main/java/net/imglib2/algorithm/morphology/BlackTopHat.java b/src/main/java/net/imglib2/algorithm/morphology/BlackTopHat.java index aabca07f8..e04715855 100644 --- a/src/main/java/net/imglib2/algorithm/morphology/BlackTopHat.java +++ b/src/main/java/net/imglib2/algorithm/morphology/BlackTopHat.java @@ -265,6 +265,7 @@ public static < T extends Type< T > & Comparable< T > & Sub< T > > Img< T > blac * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * The structuring element is specified through a list of {@link Shape}s, to - * allow for performance optimization through structuring element - * decomposition. Each shape is processed in order as given in the list. If - * the list is empty, the target receives a copy of the source. - *
- * This method differs from - * {@link #blackTopHat(RandomAccessible, IterableInterval, List, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.blackTopHat', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strels - * the list of {@link Shape}s that serves as a structuring - * element. - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * The closing operation is simply a dilation followed by an erosion. - *
- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * The structuring element is specified through a list of {@link Shape}s, to - * allow for performance optimization through structuring element - * decomposition. Each shape is processed in order as given in the list. If - * the list is empty, the target is left untouched. - *
- * This method differs from - * {@link #close(RandomAccessible, IterableInterval, List, int) - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.close', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strels - * the structuring element, as a list of {@link Shape}s. - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * The closing operation is simply a dilation followed by an erosion. - *
- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * The structuring element is specified through a list of {@link Shape}s, to - * allow for performance optimization through structuring element - * decomposition. Each shape is processed in order as given in the list. If - * the list is empty, the target is left untouched. - *
- * This method relies on specified minimal and maximal values to start - * comparing to other pixels in the neighborhood. For this code to properly - * perform closing, it is sufficient that the specified max value is larger - * (against {@link Comparable}) than any of the value found in the source - * image, and conversely for the min value. These normally unseen parameters - * are required to operate on - * {@code T extends Comparable & Type}. - *
- * This method differs from - * {@link #close(RandomAccessible, IterableInterval, Shape, Type, Type, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.close', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strels - * the structuring element, as a list of {@link Shape}s. - * @param minVal - * a T containing set to a value smaller than any of the values - * in the source (against {@link Comparable}. - * @param maxVal - * a T containing set to a value larger than any of the values in - * the source (against {@link Comparable}. - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * The closing operation is simply a dilation followed by an erosion. - *
- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * This method differs from - * {@link #close(RandomAccessible, IterableInterval, Shape, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.close', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strel - * the {@link Shape} that serves as a structuring element. - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * The closing operation is simply a dilation followed by an erosion. - *
- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * This method relies on specified minimal and maximal values to start - * comparing to other pixels in the neighborhood. For this code to properly - * perform closing, it is sufficient that the specified max value is larger - * (against {@link Comparable}) than any of the value found in the source - * image, and conversely for the min value. These normally unseen parameters - * are required to operate on - * {@code T extends Comparable & Type}. - *
- * This method differs from - * {@link #close(RandomAccessible, IterableInterval, Shape, Type, Type, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.close', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strel - * the {@link Shape} that serves as a structuring element. - * @param minVal - * a T containing set to a value smaller than any of the values - * in the source (against {@link Comparable}. - * @param maxVal - * a T containing set to a value larger than any of the values in - * the source (against {@link Comparable}. - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * The structuring element is specified through a list of {@link Shape}s, to - * allow for performance optimization through structuring element - * decomposition. Each shape is processed in order as given in the list. If - * the list is empty, the target is left untouched. - *
- * This method differs from - * {@link #dilate(RandomAccessible, IterableInterval, List, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.dilate', type=Computer - * @param source - * the source {@link RandomAccessible}, must be sufficiently - * padded. - * @param strels - * the structuring element, as a list of {@link Shape}s. - * @param numThreads - * the number of threads to use for the calculation. - * @param target - * the target image. - */ - public static < T extends RealType< T >> void dilate( final RandomAccessible< T > source, final List< ? extends Shape > strels, final int numThreads, final IterableInterval< T > target ) - { - dilate( source, target, strels, numThreads ); - } - /** * Performs the dilation morphological operation, using a * {@link RandomAccessible} as a source and writing results in an @@ -367,6 +321,7 @@ public static < T extends RealType< T >> void dilate( final RandomAccessible< T * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * The structuring element is specified through a list of {@link Shape}s, to - * allow for performance optimization through structuring element - * decomposition. Each shape is processed in order as given in the list. If - * the list is empty, the target is left untouched. - *
- * This method relies on a specified minimal value to start comparing to - * other pixels in the neighborhood. For this code to properly perform - * dilation, it is sufficient that the specified min value is smaller - * (against {@link Comparable}) than any of the value found in the source - * image. This normally unseen parameter is required to operate on - * {@code T extends Comparable & Type}. - *
- * This method differs from - * {@link #dilate(RandomAccessible, IterableInterval, List, Type, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.dilate', type=Computer - * @param source - * the source {@link RandomAccessible}, must be sufficiently - * padded. - * @param strels - * the structuring element, as a list of {@link Shape}s. - * @param minVal - * a T containing set to a value smaller than any of the values - * in the source (against {@link Comparable}. This is required to - * perform a proper mathematical dilation. Because we operate on - * a generic {@link Type}, it has to be provided manually. - * @param numThreads - * the number of threads to use for the calculation. - * @param target - * the target image. - * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * This method differs from - * {@link #dilate(RandomAccessible, IterableInterval, Shape, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.dilate', type=Computer - * @param source - * the source {@link RandomAccessible}, must be sufficiently - * padded. - * @param strel - * the structuring element, as a {@link Shape}. - * @param numThreads - * the number of threads to use for the calculation. - * @param target - * the target image. - */ - public static < T extends RealType< T >> void dilate( final RandomAccessible< T > source, final Shape strel, final int numThreads, final IterableInterval< T > target ) - { - dilate( source, target, strel, numThreads ); - } - /** * Performs the dilation morphological operation, using a * {@link RandomAccessible} as a source and writing results in an @@ -601,6 +453,7 @@ public static < T extends RealType< T >> void dilate( final RandomAccessible< T * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * This method relies on a specified minimal value to start comparing to - * other pixels in the neighborhood. For this code to properly perform - * dilation, it is sufficient that the specified min value is smaller - * (against {@link Comparable}) than any of the value found in the source - * image. This normally unseen parameter is required to operate on - * {@code T extends Comparable & Type}. - *
- * This method differs from - * {@link #dilate(RandomAccessible, IterableInterval, Shape, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.dilate', type=Computer - * @param source - * the source {@link RandomAccessible}, must be sufficiently - * padded. - * @param strel - * the structuring element, as a {@link Shape}. - * @param minVal - * a T containing set to a value smaller than any of the values - * in the source (against {@link Comparable}. This is required to - * perform a proper mathematical dilation. Because we operate on - * a generic {@link Type}, it has to be provided manually. - * @param numThreads - * the number of threads to use for the calculation. - * @param target - * the target image. - * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * The structuring element is specified through a list of {@link Shape}s, to - * allow for performance optimization through structuring element - * decomposition. Each shape is processed in order as given in the list. If - * the list is empty, the target is left untouched. - *
- * This method differs from - * {@link #erode(RandomAccessible, IterableInterval, Shape, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.erode', type=Computer - * @param source - * the source {@link RandomAccessible}, must be sufficiently - * padded. - * @param target - * the target image. - * @param strels - * the structuring element, as a list of {@link Shape}s. - * @param numThreads - * the number of threads to use for the calculation. - */ - public static < T extends RealType< T >> void erode( final RandomAccessible< T > source, final List< ? extends Shape > strels, final int numThreads, final IterableInterval< T > target ){ - erode(source, target, strels, numThreads); - } - /** * Performs the erosion morphological operation, using a * {@link RandomAccessible} as a source and writing results in an @@ -366,6 +320,7 @@ public static < T extends RealType< T >> void erode( final RandomAccessible< T > * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * The structuring element is specified through a list of {@link Shape}s, to - * allow for performance optimization through structuring element - * decomposition. Each shape is processed in order as given in the list. If - * the list is empty, the target is left untouched. - *
- * This method relies on a specified maximal value to start comparing to - * other pixels in the neighborhood. For this code to properly perform - * erosion, it is sufficient that the specified max value is larger (against - * {@link Comparable}) than any of the value found in the source image. This - * normally unseen parameter is required to operate on - * {@code T extends Comparable & Type}. - *
- * This method differs from - * {@link #erode(RandomAccessible, IterableInterval, List, Type, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.erode', type=Computer - * @param source - * the source {@link RandomAccessible}, must be sufficiently - * padded. - * @param strels - * the structuring element, as a list of {@link Shape}s. - * @param maxVal - * a T containing set to a value larger than any of the values in - * the source (against {@link Comparable}. This is required to - * perform a proper mathematical erosion. Because we operate on a - * generic {@link Type}, it has to be provided manually. - * @param numThreads - * the number of threads to use for the calculation. - * @param target - * the target image. - * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * This method differs from - * {@link #erode(RandomAccessible, IterableInterval, Shape, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.erode', type=Computer - * @param source - * the source {@link RandomAccessible}, must be sufficiently - * padded. - * @param strel - * the structuring element, as a {@link Shape}. - * @param numThreads - * the number of threads to use for the calculation. - * @param target - * the target image. - */ - public static < T extends RealType< T >> void erode( final RandomAccessible< T > source, final Shape strel, final int numThreads, final IterableInterval< T > target ){ - erode(source, target, strel, numThreads); - } - /** * Performs the erosion morphological operation, using a * {@link RandomAccessible} as a source and writing results in an @@ -598,6 +452,7 @@ public static < T extends RealType< T >> void erode( final RandomAccessible< T > * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * This method relies on a specified maximal value to start comparing to - * other pixels in the neighborhood. For this code to properly perform - * erosion, it is sufficient that the specified max value is larger (against - * {@link Comparable}) than any of the value found in the source image. This - * normally unseen parameter is required to operate on - * {@code T extends Comparable & Type}. - *
- * This method differs from - * {@link #erode(RandomAccessible, IterableInterval, Shape, Type, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.erode', type=Computer - * @param source - * the source {@link RandomAccessible}, must be sufficiently - * padded. - * @param strel - * the structuring element, as a {@link Shape}. - * @param maxVal - * a T containing set to a value larger than any of the values in - * the source (against {@link Comparable}. This is required to - * perform a proper mathematical erosion. Because we operate on a - * generic {@link Type}, it has to be provided manually. - * @param numThreads - * the number of threads to use for the calculation. - * @param target - * the target image. - * @param- * The opening operation is simply an erosion followed by a dilation. - *
- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * The structuring element is specified through a list of {@link Shape}s, to - * allow for performance optimization through structuring element - * decomposition. Each shape is processed in order as given in the list. If - * the list is empty, the target is left untouched. - *
- * This method differs from - * {@link #open(RandomAccessible, IterableInterval, List, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.erode', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strels - * the structuring element, as a list of {@link Shape}s. - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * The opening operation is simply an erosion followed by a dilation. - *
- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * The structuring element is specified through a list of {@link Shape}s, to - * allow for performance optimization through structuring element - * decomposition. Each shape is processed in order as given in the list. If - * the list is empty, the target is left untouched. - *
- * This method relies on specified minimal and maximal values to start - * comparing to other pixels in the neighborhood. For this code to properly - * perform opening, it is sufficient that the specified max value is larger - * (against {@link Comparable}) than any of the value found in the source - * image, and conversely for the min value. These normally unseen parameters - * are required to operate on - * {@code T extends Comparable & Type}. - *
- * This method differs from - * {@link #open(RandomAccessible, IterableInterval, List, Type, Type, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.erode', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strels - * the structuring element, as a list of {@link Shape}s. - * @param minVal - * a T containing set to a value smaller than any of the values - * in the source (against {@link Comparable}). - * @param maxVal - * a T containing set to a value larger than any of the values in - * the source (against {@link Comparable}). - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * The opening operation is simply an erosion followed by a dilation. - *
- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * This method differs from - * {@link #open(RandomAccessible, IterableInterval, Shape, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.open', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strel - * the {@link Shape} that serves as a structuring element. - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * The opening operation is simply an erosion followed by a dilation. - *
- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * This method relies on specified minimal and maximal values to start - * comparing to other pixels in the neighborhood. For this code to properly - * perform opening, it is sufficient that the specified max value is larger - * (against {@link Comparable}) than any of the value found in the source - * image, and conversely for the min value. These normally unseen parameters - * are required to operate on - * {@code T extends Comparable & Type}. - *
- * This method differs from - * {@link #open(RandomAccessible, IterableInterval, Shape, Type, Type, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.erode', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strel - * the {@link Shape} that serves as a structuring element. - * @param minVal - * a T containing set to a value smaller than any of the values - * in the source (against {@link Comparable}). - * @param maxVal - * a T containing set to a value larger than any of the values in - * the source (against {@link Comparable}). - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * The structuring element is specified through a list of {@link Shape}s, to - * allow for performance optimization through structuring element - * decomposition. Each shape is processed in order as given in the list. If - * the list is empty, the target receives a copy of the source. - *
- * This method differs from - * {@link #topHat(RandomAccessible, IterableInterval, List, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.topHat, morphology.whiteTopHat', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strels - * the list of {@link Shape}s that serves as a structuring - * element. - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * The structuring element is specified through a list of {@link Shape}s, to - * allow for performance optimization through structuring element - * decomposition. Each shape is processed in order as given in the list. If - * the list is empty, the target receives a copy of the source. - *
- * This method relies on specified minimal and maximal values to start - * comparing to other pixels in the neighborhood. For this code to properly - * perform, it is sufficient that the specified max value is larger (against - * {@link Comparable}) than any of the value found in the source image, and - * conversely for the min value. These normally unseen parameters are - * required to operate on - * {@code T extends Comparable & Sub}. - *
- * This method differs from - * {@link #topHat(RandomAccessible, IterableInterval, List, Type, Type, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.topHat, morphology.whiteTopHat', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strels - * the list of {@link Shape}s that serves as a structuring - * element. - * @param minVal - * a T containing set to a value smaller than any of the values - * in the source (against {@link Comparable}). - * @param maxVal - * a T containing set to a value larger than any of the values in - * the source (against {@link Comparable}). - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * This method differs from - * {@link #topHat(RandomAccessible, IterableInterval, Shape, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.topHat, morphology.whiteTopHat', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strel - * the {@link Shape} that serves as a structuring element. - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * Careful: Target must point to a different structure than source. - * In place operation will not work but will not generate an error. - *
- * It is the caller responsibility to ensure that the source is sufficiently - * padded to properly cover the target range plus the shape size. See - * e.g. {@link Views#extendValue(RandomAccessibleInterval, Type)} - *
- * It is limited to flat structuring elements, only having - * {@code on/off} pixels, contrary to grayscale structuring elements. - * This allows to simply use a {@link Shape} as a type for these structuring - * elements. - *
- * This method relies on specified minimal and maximal values to start - * comparing to other pixels in the neighborhood. For this code to properly - * perform, it is sufficient that the specified max value is larger (against - * {@link Comparable}) than any of the value found in the source image, and - * conversely for the min value. These normally unseen parameters are - * required to operate on - * {@code T extends Comparable & Sub}. - *
- * This method differs from - * {@link #topHat(RandomAccessible, IterableInterval, Shape, Type, Type, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='morphology.topHat, morphology.whiteTopHat', type=Computer - * @param source - * the {@link RandomAccessible} to operate on. - * @param strel - * the {@link Shape} that serves as a structuring element. - * @param minVal - * a T containing set to a value smaller than any of the values - * in the source (against {@link Comparable}). - * @param maxVal - * a T containing set to a value larger than any of the values in - * the source (against {@link Comparable}). - * @param numThreads - * the number of threads to use for calculation. - * @param target - * the {@link IterableInterval} to write the results on. - * @param- * This method differs from - * {@link #transform(RandomAccessible, RandomAccessibleInterval, DISTANCE_TYPE, double...)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.distanceTransform', type=Computer - * @param source - * Input function on which distance transform should be computed. - * @param distanceType - * Defines distance to be used: squared Euclidian or L1 - * @param weights - * Individual weights for each dimension, balancing image values - * and distance (when using squared Euclidian distance, weights - * should be squared, too). - * @param target - * Intermediate and final results of distance transform. - * @param- * This method differs from - * {@link #transform(RandomAccessible, RandomAccessibleInterval, RandomAccessibleInterval, DISTANCE_TYPE, double...)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.distanceTransform', type=Computer - * @param source - * Input function on which distance transform should be computed. - * @param tmp - * Storage for intermediate results. - * @param distanceType - * Defines distance to be used: squared Euclidian or L1 - * @param weights - * Individual weights for each dimension, balancing image values - * and distance (when using squared Euclidian distance, weights - * should be squared, too). - * @param target - * Final result of distance transform. - * @param- * This method differs from - * {@link #transform(RandomAccessible, RandomAccessibleInterval, RandomAccessibleInterval, DISTANCE_TYPE, ExecutorService, int, double...)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @param source - * Input function on which distance transform should be computed. - * @param tmp - * Storage for intermediate results. - * @param distanceType - * Defines distance to be used: squared Euclidian or L1 - * @param es - * {@link ExecutorService} for parallel execution. - * @param nTasks - * Number of tasks/parallelism - * @param weights - * Individual weights for each dimension, balancing image values - * and distance (when using squared Euclidian distance, weights - * should be squared, too). - * @param target - * Final result of distance transform. - * @param- * This method differs from - * {@link #transform(RandomAccessible, RandomAccessibleInterval, Distance)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @param source - * Input function on which distance transform should be computed. - * @param d - * {@link Distance} between two points. - * @param target - * Final result of distance transform. - * @param- * This method differs from - * {@link #transform(RandomAccessible, RandomAccessibleInterval, Distance, ExecutorService, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.distanceTransform', type=Computer - * @param source - * Input function on which distance transform should be computed. - * @param d - * {@link Distance} between two points. - * @param es - * {@link ExecutorService} for parallel execution. - * @param nTasks - * Number of tasks/parallelism - * @param target - * Final result of distance transform. - * @param- * This method differs from - * {@link #transform(RandomAccessible, RandomAccessibleInterval, RandomAccessibleInterval, Distance)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
* - * @implNote op name='image.distanceTransform', type=Computer * @param source * Input function on which distance transform should be computed. * @param tmp * Storage for intermediate results. - * @param d - * {@link Distance} between two points. * @param target * Final result of distance transform. + * @param d + * {@link Distance} between two points. + * @param es + * {@link ExecutorService} for parallel execution. + * @param nTasks + * Number of tasks/parallelism * @param- * This method differs from - * {@link #transform(RandomAccessible, RandomAccessibleInterval, RandomAccessibleInterval, Distance, ExecutorService, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @param source - * Input function on which distance transform should be computed. - * @param tmp - * Storage for intermediate results. - * @param d - * {@link Distance} between two points. - * @param es - * {@link ExecutorService} for parallel execution. - * @param nTasks - * Number of tasks/parallelism - * @param target - * Final result of distance transform. - * @param- * This method differs from - * {@link #binaryTransform(RandomAccessible, RandomAccessibleInterval, DISTANCE_TYPE, double...)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.binaryDistanceTransform', type=Computer - * @param source - * Input function on which distance transform should be computed. - * @param distanceType - * Defines distance to be used: squared Euclidian or L1 - * @param weights - * Individual weights for each dimension, balancing image values - * and distance (when using squared Euclidian distance, weights - * should be squared, too). - * @param target - * Intermediate and final results of distance transform. - * @param - * {@link BooleanType} binary mask input - * @param - * {@link RealType} intermediate results - */ - public static < B extends BooleanType< B >, U extends RealType< U > > void binaryTransform( - final RandomAccessible< B > source, - final DISTANCE_TYPE distanceType, - final double[] weights, - final RandomAccessibleInterval< U > target ) - { - binaryTransform( source, target, distanceType, weights ); - } - /** * Create binary distance transform on {@code source} using squared * Euclidian (L2) or L1 distance. Intermediate results will be stored in @@ -1088,6 +701,7 @@ public static < B extends BooleanType< B >, U extends RealType< U > > void binar * @throws ExecutionException * if the computation threw an exception (distance transform may * be computed only partially) + * @implNote op name='image.distanceTransform', priority='100', type=Computer2 */ public static < B extends BooleanType< B >, U extends RealType< U > > void binaryTransform( final RandomAccessible< B > source, @@ -1100,57 +714,6 @@ public static < B extends BooleanType< B >, U extends RealType< U > > void binar binaryTransform( source, target, target, distanceType, es, nTasks, weights ); } - /** - * Create binary distance transform on {@code source} using squared - * Euclidian (L2) or L1 distance. Intermediate results will be stored in - * {@code target} ({@link DoubleType} recommended). The distance can be - * weighted (individually for each dimension, if desired) against the image - * values via the weights parameter. - *- * This method differs from - * {@link #binaryTransform(RandomAccessible, RandomAccessibleInterval, DISTANCE_TYPE, ExecutorService, int, double...)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.binaryDistanceTransform', type=Computer - * @param source - * Input function on which distance transform should be computed. - * @param distanceType - * Defines distance to be used: squared Euclidian or L1 - * @param es - * {@link ExecutorService} for parallel execution. - * @param nTasks - * Number of tasks/parallelism - * @param weights - * Individual weights for each dimension, balancing image values - * and distance (when using squared Euclidian distance, weights - * should be squared, too). - * @param target - * Intermediate and final results of distance transform. - * @param - * {@link BooleanType} binary mask input - * @param - * {@link RealType} intermediate results - * @throws InterruptedException - * if interrupted while waiting, in which case unfinished tasks - * are cancelled (distance transform may be computed only - * partially) - * @throws ExecutionException - * if the computation threw an exception (distance transform may - * be computed only partially) - */ - public static < B extends BooleanType< B >, U extends RealType< U > > void binaryTransform( - final RandomAccessible< B > source, - final DISTANCE_TYPE distanceType, - final ExecutorService es, - final int nTasks, - final double[] weights, - final RandomAccessibleInterval< U > target ) throws InterruptedException, ExecutionException - { - binaryTransform( source, target, distanceType, es, nTasks, weights ); - } - /** * Create binary distance transform on {@code source} using squared * Euclidian (L2) or L1 distance. Intermediate results will be stored in @@ -1177,6 +740,7 @@ public static < B extends BooleanType< B >, U extends RealType< U > > void binar * {@link RealType} intermediate results * @param- * This method differs from - * {@link #binaryTransform(RandomAccessible, RandomAccessibleInterval, RandomAccessibleInterval, DISTANCE_TYPE, double...)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.binaryDistanceTransform', type=Computer - * @param source - * Input function on which distance transform should be computed. - * @param tmp - * Storage for intermediate results. - * @param distanceType - * Defines distance to be used: squared Euclidian or L1 - * @param weights - * Individual weights for each dimension, balancing image values - * and distance (when using squared Euclidian distance, weights - * should be squared, too). - * @param target - * Final result of distance transform. - * @param - * {@link BooleanType} binary mask input - * @param - * {@link RealType} intermediate results - * @param- * This method differs from - * {@link #binaryTransform(RandomAccessible, RandomAccessibleInterval, RandomAccessibleInterval, DISTANCE_TYPE, ExecutorService, int, double...)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.binaryDistanceTransform', type=Computer - * @param source - * Input function on which distance transform should be computed. - * @param tmp - * Storage for intermediate results. - * @param distanceType - * Defines distance to be used: squared Euclidian or L1 - * @param es - * {@link ExecutorService} for parallel execution. - * @param nTasks - * Number of tasks/parallelism - * @param weights - * Individual weights for each dimension, balancing image values - * and distance (when using squared Euclidian distance, weights - * should be squared, too). - * @param target - * Final result of distance transform. - * @param - * {@link BooleanType} binary mask input - * @param - * {@link RealType} intermediate results - * @param- * This method differs from - * {@link #binaryTransform(RandomAccessible, RandomAccessibleInterval, Distance) - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.binaryDistanceTransform', type=Computer - * @param source - * Input function on which distance transform should be computed. - * @param d - * {@link Distance} between two points. - * @param target - * Final result of distance transform. - * @param - * {@link BooleanType} binary mask input - * @param - * {@link RealType} intermediate results - */ - public static < B extends BooleanType< B >, U extends RealType< U > > void binaryTransform( - final RandomAccessible< B > source, - final Distance d, - final RandomAccessibleInterval< U > target ) - { - binaryTransform( source, target, d ); - } - /** * Create binary distance transform on {@code source} using arbitrary * {@link Distance} d. Intermediate and final results will be stored in @@ -1462,6 +896,7 @@ public static < B extends BooleanType< B >, U extends RealType< U > > void binar * @throws ExecutionException * if the computation threw an exception (distance transform may * be computed only partially) + * @implNote op name='image.distanceTransform', priority='100', type=Computer2 */ public static < B extends BooleanType< B >, U extends RealType< U > > void binaryTransform( final RandomAccessible< B > source, @@ -1473,50 +908,6 @@ public static < B extends BooleanType< B >, U extends RealType< U > > void binar binaryTransform( source, target, target, d, es, nTasks ); } - /** - * Create binary distance transform on {@code source} using arbitrary - * {@link Distance} d. Intermediate and final results will be stored in - * {@code target} ({@link DoubleType} recommended). - *- * This method differs from - * {@link #binaryTransform(RandomAccessible, RandomAccessibleInterval, Distance, ExecutorService, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.binaryDistanceTransform', type=Computer - * @param source - * Input function on which distance transform should be computed. - * @param target - * Final result of distance transform. - * @param d - * {@link Distance} between two points. - * @param es - * {@link ExecutorService} for parallel execution. - * @param nTasks - * Number of tasks/parallelism - * @param - * {@link BooleanType} binary mask input - * @param - * {@link RealType} intermediate results - * @throws InterruptedException - * if interrupted while waiting, in which case unfinished tasks - * are cancelled (distance transform may be computed only - * partially) - * @throws ExecutionException - * if the computation threw an exception (distance transform may - * be computed only partially) - */ - public static < B extends BooleanType< B >, U extends RealType< U > > void binaryTransform( - final RandomAccessible< B > source, - final Distance d, - final ExecutorService es, - final int nTasks, - final RandomAccessibleInterval< U > target) throws InterruptedException, ExecutionException - { - binaryTransform( source, target, d, es, nTasks ); - } - /** * Create binary distance transform on {@code source} using arbitrary * {@link Distance} d. Intermediate results will be stored in {@code tmp} @@ -1537,6 +928,7 @@ public static < B extends BooleanType< B >, U extends RealType< U > > void binar * {@link RealType} intermediate results * @param- * This method differs from - * {@link #binaryTransform(RandomAccessible, RandomAccessibleInterval, RandomAccessibleInterval, Distance)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.binaryDistanceTransform', type=Computer - * @param source - * Input function on which distance transform should be computed. - * @param tmp - * Storage for intermediate results. - * @param target - * Final result of distance transform. - * @param d - * {@link Distance} between two points. - * @param - * {@link BooleanType} binary mask input - * @param - * {@link RealType} intermediate results - * @param- * This method differs from - * {@link #binaryTransform(RandomAccessible, RandomAccessibleInterval, RandomAccessibleInterval, Distance, ExecutorService, int)} - * only in that its parameter order is tailored to an Op. The output comes - * last, and the primary input (the input image) comes first. - *
- * - * @implNote op name='image.binaryDistanceTransform', type=Computer - * @param source - * Input function on which distance transform should be computed. - * @param tmp - * Storage for intermediate results. - * @param d - * {@link Distance} between two points. - * @param es - * {@link ExecutorService} for parallel execution. - * @param nTasks - * Number of tasks/parallelism - * @param target - * Final result of distance transform. - * @param - * {@link BooleanType} binary mask input - * @param - * {@link RealType} intermediate results - * @param