Hi all,
I would like to discuss the configuration of the N4 bias field correction step in the t1-linear pipeline, specifically regarding the parameter:
n4biascorrection.inputs.bspline_fitting_distance = 600
|
n4biascorrection.inputs.bspline_fitting_distance = 600 |
According to the ANTs documentation https://github.com/ANTsX/ANTs/wiki/N4BiasFieldCorrection, the spline spacing (-b) parameter should typically be set within a certain range for human brain MRI:
“For human brain images, this should generally be somewhere between 100 and 200 mm.”
Additionally, in ANTsPy / ANTsR, the default spline spacing is around: https://antsx.github.io/ANTsR/reference/n4BiasFieldCorrection.html
“Default … is 200 mm per mesh element in each dimension.”
Would it be reasonable to modify this parameter ? (e.g., to 200 mm) for:
- better computational efficiency
- potentially improved bias correction fidelity
CHEERS～
Hi all,
I would like to discuss the configuration of the N4 bias field correction step in the t1-linear pipeline, specifically regarding the parameter:
n4biascorrection.inputs.bspline_fitting_distance = 600
clinica/clinica/pipelines/t1_linear/anat_linear_pipeline.py
Line 265 in abb12cb
According to the ANTs documentation https://github.com/ANTsX/ANTs/wiki/N4BiasFieldCorrection, the spline spacing (-b) parameter should typically be set within a certain range for human brain MRI:
“For human brain images, this should generally be somewhere between 100 and 200 mm.”
Additionally, in ANTsPy / ANTsR, the default spline spacing is around: https://antsx.github.io/ANTsR/reference/n4BiasFieldCorrection.html
“Default … is 200 mm per mesh element in each dimension.”
Would it be reasonable to modify this parameter ? (e.g., to 200 mm) for:
CHEERS～