installation.rst

Installation

Note

To follow these instructions locally you will need a local copy of this package. Start by cloning this project with something like:

$ git clone https://github.com/bioidiap/bob.example.faceverify.git
$ cd bob.example.faceverify

Installation of the toolkit uses the buildout build environment. You don't need to understand its inner workings to use this package. Here is a recipe to get you started (shell commands are marked with a $ signal):

$ python bootstrap.py
$ ./bin/buildout

These 2 commands should download and install all non-installed dependencies and get you a fully operational test and development environment.

Note

The python shell used in the first line of the previous command set determines the python interpreter that will be used for all scripts developed inside this package. Because this package makes use of Bob, you must make sure that the bootstrap.py script is called with the same interpreter used to build Bob, or unexpected problems might occur.

If Bob is installed by the administrator of your system, it is safe to consider it uses the default python interpreter. In this case, the above 3 command lines should work as expected.

Use this example with Bob not installed globally

If your Bob version is not installed globally, this package provides a configuration file localbob.cfg that has to be modified. Please search for the bob.install.directory region and set the value according to your local Bob install directory. If you are at Idiap, you can simply choose one the existing directories.

Use Bob at Idiap

To get the example running nicely at Idiap, as noted above, bootstrap.py has to be executed with the correct python version. For Idiap, this is (currently):

$ /idiap/group/torch5spro/nightlies/externals/v2/linux-x86_64/bin/python2.6 bootstrap.py
$ ./bin/buildout -c localbob.cfg

Downloading the test database

The images that are required to run the test are not included in this package, but they are freely downloadable from the internet:

http://www.cl.cam.ac.uk/research/dtg/attarchive/facedatabase.html

Unpack the database in a directory that fits you. The easiest solution is to create a subdirectory Database in this package. If you decide to put the data somewhere else, please remember the image directory.

Note

If you are at Idiap, the AT&T database is located at /idiap/group/biometric/databases/orl. To ease up the usage of the examples, you can generate a link to the database:

$ ln -s /idiap/group/biometric/databases/orl Database