How can I represent avi video as set of matrices using Python?

Question

I have video files written in avi format and I would like to analyze these videos using Python. For that I would like to represent every frame of the video as a 2D matrix.

How can I do that? Google search gives me PyMedia as a way to go? Is it really the best choice or there some other approaches that I should to considered?

If the PyMedia is a good choice, could anybody pleas to give me a link where I can get exe files to install the module on Windows from binaries?

By the way, is it a good idea, in general, to use Python for these purposes? I like Python very much because of its simplicity and I prefer to use it, but if it is really not suitable for analysis of video, I am ready to use something else.

ADDED Some people claim that PyMedia is "dead". Is it true?

Answer 1

Yeah, the latest news on the PyMedia web site is dated 01 Feb 2006. That's a pretty bad sign.

The most active and up-to-date open project for manipulating video is ffmpeg . Apparently there is a recently updated python wrapper for it: http://code.google.com/p/pyffmpeg/

In general Python is much too slow for doing any sort of pixel analysis of video. Therefore there will be practically zero libraries of any reasonable level of quality and support for helping at the pixel level of granularity. There are well supported libraries for working at an image level of granularity though. PIL seems to be a popular choice: http://www.pythonware.com/products/pil/