Programming Computer Vision with Python

Appendix A. Installing Packages

Here are short installation instructions for the packages used in the book. They are written based on the latest versions as of the writing of this book. Things change (URLs change!), so if the instructions become outdated, check the individual project websites for help.

In addition to the specific instructions, an option that often works on most platforms is Python’s easy_install. If you run into problems with the installation instructions given here, easy_install is worth a try. Find out more on the package website, http://packages.python.org/distribute/easy_install.html.

A.1 NumPy and SciPy

Installing NumPy and SciPy is a little different depending on your operating system. Follow the applicable instructions below. The current versions are 2.0 (NumPy) and 0.11 (SciPy) on most platforms. A package that currently works on all major platforms is the Enthought EPD Free bundle, a free light version of the commercial Enthought distribution, available at http://enthought.com/products/epd_free.php.

Windows

The easiest way to install NumPy and SciPy is to download and install the binary distributions from http://www.scipy.org/Download.

Mac OS X

Later versions of Mac OS X (10.7.0 [Lion] and up) come with NumPy pre-installed.

An easy way to install NumPy and SciPy for Mac OS X is with the “superpack” from https://github.com/fonnesbeck/ScipySuperpack. This also gives you Matplotlib.

Another alternative is to use the package system MacPorts (http://www.macports.org/ ...