T
- public class DefaultCreateKernelBiGauss<T extends ComplexType<T>> extends AbstractBinaryFunctionOp<double[],Integer,RandomAccessibleInterval<T>> implements Ops.Create.KernelBiGauss
The BiGauss kernel is a composition of two standard Gauss kernels. If we were to assume 1D kernel centered at zero (0), an inner kernel of the BiGauss, with its shape given with sigmas[0], would span from -sigmas[0] till +sigmas[0]; outer kernel, with its shape given with sigmas[1], surrounds the inner, e.g. for the positive side, from sigmas[0] till sigmas[0]+2*sigmas[1] and its center having at sigmas[0]-sigmas[1]. That is, the inner Gauss exist up to its inflection points from which the filter takes shape of the outer Gauss. Both kernels are, however, appropriately scaled and shifted to obtain a smooth BiGauss kernel.
Note that the kernel is always isotropic. The second parameter gives dimensionality of the created kernel.
All values are in units of pixels.
Literature:C. Xiao, M. Staring, Y. Wang, D.P. Shamonin, B.C. Stoel. Multiscale Bi-Gaussian Filter for Adjacent Curvilinear Structures Detection with Application to Vasculature Images. IEEE TMI, vol. 22, no. 1, 2013.
SpecialOp.Flavor
NAME
Constructor and Description |
---|
DefaultCreateKernelBiGauss() |
Modifier and Type | Method and Description |
---|---|
RandomAccessibleInterval<T> |
calculate(double[] sigmas,
Integer dimensionality)
Calculates the output given two inputs.
|
void |
initialize()
Initializes the object.
|
in1, in2, out, run, setInput1, setInput2
ops, setEnvironment
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
ops, setEnvironment
calculate, getIndependentInstance, run
in, setInput
calculate, run
run
candidates, filterArity, op
public void initialize()
Initializable
initialize
in interface Initializable
public RandomAccessibleInterval<T> calculate(double[] sigmas, Integer dimensionality)
BinaryFunctionOp
calculate
in interface BinaryFunctionOp<double[],Integer,RandomAccessibleInterval<T extends ComplexType<T>>>
sigmas
- first argument to the functiondimensionality
- second argument to the functionCopyright © 2014–2022 ImageJ. All rights reserved.