fillShells: Assign image volumes to shells

In tractor.base: Read, Manipulate and Visualise Magnetic Resonance Images

Assign image volumes to shells

Description

This function guesses the assignment of image volumes to shells, expanding a list of unique b-values to one value per volume. This is achieved by performing k-means clustering on the mean intensities of each volume. The set of unique b-values must be known beforehand, and no sanity checking is performed on the intensities for each shell except ordering.

Usage

fillShells(image, bValues)

Arguments

image	A 4D diffusion-weighted image, whose mean volume intensities will be used to cluster the acquisition into shells. It does not need to already have gradient direction or b-value information associated with it.
bValues	A numeric vector of unique b-values.

Value

A numeric vector expanding the original bValues to one value per volume of the image.

Note

The procedure performed by this function is a heuristic, and may produce inaccurate results if some shell b-values are close together, if there are many unique b-values, if the assumption of decreasing mean intensity with increasing b-value doesn't hold, etc. If the full acquisition scheme is known then it is alway better to specify it explicitly.

Author(s)

Jon Clayden

References

Please cite the following reference when using TractoR in your work:

J.D. Clayden, S. Muñoz Maniega, A.J. Storkey, M.D. King, M.E. Bastin & C.A. Clark (2011). TractoR: Magnetic resonance imaging and tractography with R. Journal of Statistical Software 44(8):1-18. \Sexpr[results=rd]{tools:::Rd_expr_doi("10.18637/jss.v044.i08")}.

tractor.base documentation built on Nov. 5, 2025, 6:18 p.m.