InstallLinux

Prerequisite

gcc
boost-dev
fftw3-dev
tinyXML
openCV
subversion
beautifull soup 4
numpy
scipy
numexpr
pygraph

On Debian, Ubuntu, etc. this can be done via

sudo apt-get install build-essential libboost-all-dev libfftw3-dev libtinyxml-dev libopencv-dev subversion python-bs4 python-numpy python-scipy python-numexpr python-pygraph

For the C++ implementation, you will also need CImg.h in your include path.

Setup and install

You can obtain the last version of the code from GitHub

git clone https://github.com/MathieuLeocmach/colloids.git ~/src/colloids

Optimized C++ code can be compiled via.

cd ~/src/colloids
./configure --prefix=$HOME --with-boost --with-boost_program_options
make
make install

Note that most of it is rather old code, before I knew Python, and is now superseded by much lighter python implementation where only the critical parts are compiled. However, the implementation of the Crocker and Grier tracking algorithm is still of interest.

The C++ version of the multiscale tracker can be compiled via

cd ~/src/colloids/multiscale/Release
make all

The resulting multiscale file is a program you can copy in your PATH (eg. ~/local/bin). Alternatively, you can add ~/src/colloids/multiscale/Release to your PATH environment variable.

It is much more recent code, highly optimized and required for most practical 3D tracking purposes (except if you have hundred of Gb of RAM and time to spare).

To use the python implementation, you will need to set the PYTHONPATH environment variable so that it points to ~/src/colloids/python. On Ubuntu you can edit (or create) the file ~/.pam_environement and add (or edit) the line

PYTHONPATH DEFAULT=${PYTHONPATH}:${HOME}/src/colloids/python