Constructor Detail

• DifferenceOfGaussian

  public DifferenceOfGaussian()

Method Detail

• DoG

  public static <I extends NumericType<I>,T extends NumericType<T> & NativeType<T>> void DoG(double[] sigmaSmaller,
                                                                                             double[] sigmaLarger,
                                                                                             RandomAccessible<I> input,
                                                                                             RandomAccessibleInterval<T> dog,
                                                                                             ExecutorService service)

  Compute the difference of Gaussian for the input. Input convolved with Gaussian of sigmaSmaller is subtracted from input convolved with Gaussian of sigmaLarger (where sigmaLarger > sigmaSmaller).

  Creates an appropriate temporary image and calls DoG(double[], double[], RandomAccessible, RandomAccessible, RandomAccessibleInterval, ExecutorService) .

  Parameters:

  sigmaSmaller - stddev (in every dimension) of smaller Gaussian.

  sigmaLarger - stddev (in every dimension) of larger Gaussian.

  input - the input image extended to infinity (or at least covering the same interval as the dog result image, plus borders for convolution).

  dog - the Difference-of-Gaussian result image.

  service - service providing threads for multi-threading

• DoG

  public static <I extends NumericType<I>,T extends NumericType<T> & NativeType<T>> void DoG(double[] sigmaSmaller,
                                                                                             double[] sigmaLarger,
                                                                                             RandomAccessible<I> input,
                                                                                             RandomAccessible<T> tmp,
                                                                                             RandomAccessibleInterval<T> dog,
                                                                                             ExecutorService service)

  Compute the difference of Gaussian for the input. Input convolved with Gaussian of sigmaSmaller is subtracted from input convolved with Gaussian of sigmaLarger (where sigmaLarger > sigmaSmaller).

  Parameters:

  sigmaSmaller - stddev (in every dimension) of smaller Gaussian.

  sigmaLarger - stddev (in every dimension) of larger Gaussian.

  input - the input image extended to infinity (or at least covering the same interval as the dog result image, plus borders for convolution).

  tmp - temporary image, must at least cover the same interval as the dog result image.

  dog - the Difference-of-Gaussian result image.

  service - how many threads to use for the computation.

• computeSigmas

  public static double[][] computeSigmas(double imageSigma,
                                         double minf,
                                         double[] pixelSize,
                                         double sigma1,
                                         double sigma2)

  Helper function to compute per-dimension sigmas in pixel coordinates. The parameters sigma1 and sigma2 specify desired sigmas (scale) in image coordinates. Taking into account the sigma of the input image as well as the image calibration, the resulting sigma arrays specifiy the smoothing that has to be applied to achieve the desired sigmas.

  Parameters:

  imageSigma - estimated sigma of the input image, in pixel coordinates.

  minf - multiple of the imageSigma that smoothing with the resulting sigma must at least achieve.

  pixelSize - calibration. Dimensions of a pixel in image units.

  sigma1 - desired sigma in image coordinates.

  sigma2 - desired sigma in image coordinates.

  Returns:

  double[2][numDimensions], array of two arrays contains resulting sigmas for sigma1, sigma2.

• computeMinIsotropicSigma

  public static double computeMinIsotropicSigma(double imageSigma,
                                                double minf,
                                                double[] pixelSize)

  Helper function to compute the minimum sigma that can be given to computeSigmas(double, double, double[], double, double) while still achieving isotropic smoothed images.