MP-PCA

Class: NodeMpPca

Node Icon

Apply the Marchenko Pastur denoising (Veraart et al, NeuroImage 142 (2016) 394–406)

Inputs

Image

Input time series image

Type: Image4DFloat, Required, Single

Mask

Input mask

Type: Image4DBool, Optional, Single

Outputs

Result

Denoised image

Type: Image4DFloat

Pre-filter Noise std

Estimated noise before MP-PCA filter was applied

Type: Image4DFloat

Post-filter Noise std

Estimated noise after denoising

Type: Image4DFloat

Princ. Comp.

Average number of pricinpal components used in each voxel.

Type: Image4DFloat

Settings

Default

Window Text

Set the size of the window. As a rule of thumb, the number of voxels in the window should be higher than the number of frames in the series.

Number of CPUs Text

Set the number of threads to use in the calculations. Can be an integer or “auto”, which uses N-1 CPUs, where N is the number of CPUs of the system

Batch size Number

Set the number of windows to calculate per batch.

See also

Keywords: