Field Detail

• D

  protected RandomAccessibleInterval<FloatType> D

  The diffusion tensor.

• input

  protected final RandomAccessibleInterval<T extends RealType<T>> input

  The input image, will be modified by this algorithm.

• increment

  protected Img<FloatType> increment

  This is a temporary holder were we store the increment to add to the input image at each iteration. More specifically, that is dt times the right-hand-size of the diffusion equation.

• imgFactory

  protected ImgFactory<FloatType> imgFactory

• tensorComponentDimension

  protected final int tensorComponentDimension

  The dimension to iterate over to retrieve the tensor components.

Constructor Detail

• ExplicitDiffusionScheme

  public ExplicitDiffusionScheme(Img<T> input,
                                 Img<FloatType> D)

  Contructor of ExplicitDiffusionScheme, calls contructor ExplicitDiffusionScheme(RandomAccessibleInterval<T>, RandomAccessibleInterval<FloatType>, ImgFactory<FloatType>) and uses the ImgFactory of the second Img as third parameter.

  Parameters:

  input - the input Img

  D - the diffusion tensor

• ExplicitDiffusionScheme

  public ExplicitDiffusionScheme(RandomAccessibleInterval<T> input,
                                 RandomAccessibleInterval<FloatType> D)

  Contructor of ExplicitDiffusionScheme, calls contructor ExplicitDiffusionScheme(RandomAccessibleInterval<T>, RandomAccessibleInterval<FloatType>, ImgFactory<FloatType>) and uses a default ArrayImgFactory<FloatType> or CellImgFactory<FloatType> as third parameter (depends on size of input).

  Parameters:

  input - the input Img

  D - the diffusion tensor

• ExplicitDiffusionScheme

  public ExplicitDiffusionScheme(RandomAccessibleInterval<T> input,
                                 RandomAccessibleInterval<FloatType> D,
                                 ImgFactory<FloatType> imgFactory)

Method Detail

• initDensityArray

  protected abstract float[] initDensityArray()

• initDiffusionTensorArray

  protected abstract float[][] initDiffusionTensorArray()

• process

  public boolean process()

  Execute one iteration of explicit scheme of the diffusion equation.

• getIncrement

  public RandomAccessibleInterval<FloatType> getIncrement()

  Returns:

  the increment to add to the input image at each iteration. More specifically, that is dt times the right-hand-size of the diffusion equation.

• diffusionScheme

  protected abstract float diffusionScheme(float[] U,
                                           float[][] D)

  Compute the float increment of the current location, for which is given the density neighborhood and the diffusion tensor neighborhood.

  Parameters:

  U - the density neighborhood

  D - the diffusion tensor neighborhood

  Returns:

  the increment computed from the given input

  See Also:

  yieldDensity(RandomAccess, float[]), yieldDiffusionTensor(RandomAccess, float[][])

• checkInput

  public boolean checkInput()

• setDiffusionTensor

  public void setDiffusionTensor(RandomAccessibleInterval<FloatType> D)

  Set the diffusion tensor that will be used for the diffusion process.

  The diffusion tensor must be a FloatType RandomAccessibleInterval, with specific dimensions: If the target image has N dimensions, the tensor must have N+1 dimensions with:

  • the first N dimensions size equal to the input size;
  • the N+1 dimension having a size of 3.

  The tensor stores the local diffusion intensity and orientation in the shape of a real symmetric 2x2 matrix. Along the N+1 dimension, the tensor components are ordered as:

  0. Dxx
  1. Dxy
  2. Dyy

• getDiffusionTensor

  public RandomAccessibleInterval<FloatType> getDiffusionTensor()

  Returns:

  the diffusion tensor.

  See Also:

  getDiffusionTensor()

• yieldDensity

  protected abstract void yieldDensity(RandomAccess<T> ura,
                                       float[] target)

  Iterate over a nD equivalent of 3x3 neighborhood, and collect the input values needed in that neighborhood to compute the concrete diffusion scheme.

  Parameters:

  ura - the RandomAccess

  target - the float array in which the value will be stored

• yieldDiffusionTensor

  protected abstract void yieldDiffusionTensor(RandomAccess<FloatType> dra,
                                               float[][] target)

  Iterate over a nD equivalent of 3x3 neighborhood, and collect the diffusion tensor values needed in that neighborhood to compute the concrete diffusion scheme.

  Parameters:

  dra - the RandomAccess

  target - the float array in which the value will be stored