Organization
First are a set of section which address general
topics germain across all projects. After the preliminaries are taken
care of, the software is organized first by project and then
alphabetically.
License
As a rule I use
GPL 2+ for interoperability with most other software out there, such as all of it which uses
GPL 3.
Please see the license statements in the individual sources for more
details. I merely ask you cite my software when using it as part of
published research.
Bugs
Yes, they exist. If you find one,
please let me know what's up and I'll try my best to fix it. At this
time due to copyright, etc, issues I'm not accepting code submissions
unless the code has been made public domain. Even so, please contact me
before sending code. See the
Contributing section below.
Contributing
Unfortunately, due to copyright issues I
cannot accept
submitted code which has not been put into the public domain. If you
are really eager to contribute code, please contact me directly and
we'll work something out. For ease of maintenance I would rather not
need to hunt down lots of contributors prior to things like updating
the license, ergo only public domain code. The one big exception would
be if someone is interested in helping maintain the code and especially
if they are willing to help package it for distribution on
CRAN or another appropriate locale.
Code Hosting/CVS
At
the moment all code is hosted off this website with no CVS, etc, as the
codebase is small and I'm the only maintainer. That may change if
either (1) there is substantive interest in the software under
development or (2) I suddently get swamped with people wanting to help
out. Odds are as my only current projects are in R if I move the code
anywhere it will be to
R-Forge.
Source Code
The sources for various projects follow.
R package `spect'
This R package encorporates a number of scripts which were written as part of work done for my
research with Arkady Shemyakin. This package contains a number of smaller subscripts but most depend on findDistance.
Arc Polygon Solver
This R script which takes a set of points known to lie inside a circle of known radius (such as those generated by the
SPECT Sample Generator) and returns the bounding arc polygon for the center of the distribution.
- Source
- Documentation (LaTeX/PDF)
SPECT Sample Generator
The SPECT Sample Generator is an R script which generates simulated 2-D SPECT images.
- Source
- Documentation (LaTeX/PDF)
Truncated Distribution Sampler
This R script samples from the arc polygon generated by the
Arc Polygon Solver or any other closed region in the with defined vertices and whose edges are sectors of circles with known centers.
- Source
- Documentation (LaTeX/PDF)
Find Distance
This
script is a custom distance computation script in order to allow
possible expansion beyond the existing built-in distance finctions. At
this time it only takes Euclidean distance. You will need it to run any
other script as a stand alone.
- Source
- Documentation (LaTeX/PDF)
Geometric Estimators
This
script computes some basic, geometrically intuitive estimators for the
center of the SPECT image set. These are the artithmetic mean of the
vertices of the region of non-zero likelihood and the most
likely/highest posterior vertex.
- Source
- Documentation (LaTeX/PDF)
Independant Metropolis Algorithm
A
version of the Independant Metropolis Algorithm to find the maximum
posterior Bayes estimator for the center of the SPECT Image. This is
currently under active development. It has been modified to work the
with irregular geometry of the situation being modelled.
Backward Coupling Metropolis Algorithm
A
version of the Backward Coupling Metropolis Algorithm to find the
maximum posterior Bayes estimator for the center of the SPECT Image.
This is currently planned.
Adaptive Backward Coupling Metropolis Algorithm
A version of the
Adaptive Backward Coupling Metropolis Algorithm to find the maximum posterior
Bayes estimator for the center of the SPECT Image. This is currently
planned.