spline_smooth
smooth and extrapolate data in MINC volumes using spline fitting
spline_smooth [options] <infile> <outfile>
DESCRIPTION
spline_smooth smooths a MINC volume using tensor B-splines or thin plate splines. It fits a smooth spline function to the data and can optionally output a compact field representation for later evaluation with evaluate_field. This tool is a key component of the N3 non-uniformity correction pipeline, where it is used to produce a smooth estimate of the bias field from sample points.
OPTIONS
-clobber- Overwrite an existing output file.
-noclobber- Do not overwrite an existing output file (default).
-verbose- Print progress information.
-quiet- Do not print progress information.
-lambda <val>- Set the smoothing parameter controlling the trade-off between fidelity to data and smoothness (default: 0.01).
-distance <val>- Set the distance between basis functions in mm (default: 60).
-b_spline- Use tensor B-splines for smoothing (default).
-tp_spline- Use thin plate splines for smoothing.
-mask <mask.mnc>- Specify a mask volume defining the region to be smoothed.
-output_mask <mask.mnc>- Specify a mask defining the region for output.
-extrapolate- Extrapolate the field outside the mask region.
-noextrapolate- Do not extrapolate outside the mask region (default).
-full_support- Use the full support of the basis functions.
-subsample <n>- Fit to every nth data point (default: 1).
-compact <file>- Store a compact field representation to the specified file.
-novolume- Do not write the output volume; only write the compact field.
EXAMPLES
spline_smooth -distance 100 -lambda 0.001 input.mnc smoothed.mnc
spline_smooth -mask brain_mask.mnc -compact field.fld input.mnc smoothed.mnc
AUTHOR
John G. Sled - McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University.
COPYRIGHTS
Copyright © 1998 by John G. Sled