Denoising

Class: NodePatchBasedDenoising

Node Icon

Deprecated!

Implements a denoising filter that uses iterative non-local, or semi-local, weighted averaging of image patches for image denoising. The intensity at each pixel ‘p’ gets updated as a weighted average of intensities of a chosen subset of pixels from the image.

Inputs

Image

Input image.

Type: Image4DFloat, Required, Single

Outputs

Output

Resulting image.

Type: Image4DFloat

Settings

Noise

Model Selection

Select model.

Values: NoModel, Gaussian, Rician, Poisson

Noise Sigma Number

Noise std. dev.

Fidelity Weight Number

Set the fidelity weight. This weight prevents large deviations of the denoised image from the noisy data.

Kernel Bandwidth Estimation

Use Estimation Boolean

If TRUE, the kernal bandwith will be estimated automatically.

Fraction Pixels Number

Set the fraction of voxels in the image that will be used for kernel bandwidth sigma estimation. To reduce the computational burden for computing sigma, a small random fraction of the image pixels can be used.

Multiplication Factor Number

Set the kernel bandwidth sigma multiplication factor used to modify the automatically-estimated kernel bandwidth sigma. At times, it may be desirable to modify the value of the automatically-estimated sigma. Typically, this number isn't very far from 1. Note: This is used only when Use Estimation is TRUE.

Kernel Sigma Number

Set initial kernel bandwidth estimate. Note: This is changed when Use Estimation is TRUE.

Update Frequencey Integer

Set the update frequency. An optimal bandwidth will be re-estimated based on the denoised image after every ‘n’ iterations. Defaults to 3, i.e. bandwidth updated after every 3 denoising iteration.

Filter

Number Of Iterations Integer

Set the number of denoising iterations to perform. Defaults to 3.

Number Of Sample Patches Integer

Set the number of patches to sample for each pixel.

Patch Radius Integer

Set the patch radius specified in physical coordinates. Patch radius is preferably set to an even number. Currently, only isotropic patches in physical space are allowed; patches can be anisotropic in voxel space.

Sample Varience Number

Set the variance of the domain where patches are sampled.

References

Keywords: patch, denoising, noise