adapted from
PLoS supp. material
by Hermann Cuntz
adapted from
PLoS supp. material
by Hermann Cuntz
resampling is better for a more
homogeneous quadratic diameter tapering
Maps quadratic diameter tapering on a given tree structure intree.
P and ldend are derived in
(Cuntz, Borst and Segev 2007, Theor Biol Med Model, 4:21).
P is an nx3 matrix containing the parameters to put
in the quadratic equation
y = P(1)x2 +
P(2)x + P(3).
Each single triplet corresponds to the best fit to a segment
of length ldend (nx1) vector.
When the quadratic diameter is added,
the path from each terminal to the root is compared to its closest in ldend.
Then the quadratic equation is chosen according to the index in ldend.
This is done for all paths from root to terminal point and for each node
the diameter is an average of all local diameters of all paths leading through that node.
Choosing parameters (P and ldend)
by hand here is tempting but very hard.
P and ldend depend on the total leak and the minimal diameter:
these have to be adjust by the parameters scale and offset
respectively
(see “quadfit_tree”),
Example 1:
change the scaling of the quadratic tapering (scale)
>> rtree = resample_tree (sample2_tree, 1);
>> quaddiameter_tree (rtree, scale, 1)
Example 2:
add an offset diameter value to the quadratically tapering diameter (offset)
>> quaddiameter_tree (rtree, .4, offset)
This work is licensed under a
Creative Commons Attribution-Noncommercial-Share Alike 3.0 License