Seed Growing for a 4D Array

Share:

Description

Seed growing algorithm to find voxels in a three-dimensional array according to their correlation to a seed voxel. The correlation is measured according to the fourth dimension of the array.

Usage

1
extractAIF(img, x, y, z, thresh = 0.9)

Arguments

img

is the four-dimensional array of medical imaging data.

x,y,z

are the coordinates of the seed voxel.

thresh

is the minimum correlation for inclusion in the region.

Details

Correlation coefficients are computed for every voxel in the input array. A recursive algorithm is then used to grow the region of interest (ROI) from the seed voxel in three dimensions. All adjacent voxels, where the correlation exceeds the threshold, are included.

Value

coord

is a matrix of the three-dimesional coordinates (x,y,z) for all voxels found by the algorithm.

conc

is a matrix whose rows correspond to the voxels found by the algorithm and whose columns are the fourth dimension from the input array (e.g., contrast agent concentration time curve).

mask

is an array of boolean values, where only voxels included by the algorithm are given a value greater than zero.

cor

is an array that mimics the mask, but contains the estimated correlation coefficients for all voxels included by the algorithm.

Author(s)

Volker Schmid volker.schmid@users.sourceforge.net