having opencv as a dependency has two important drawbacks for me. First,
it makes the installation of the scikit more difficult. Second, the
documentation of openCV does not match the standards of numpy/scipy
docstrings (no examples, few references to articles).
Performance is great, but I think we'll gain more users by putting effort
on user-friendliness than if we put it on performance. For examples, a
large fraction of my colleagues are using image processing, but on
Windows, and they get discouraged by a package with unsufficient
I'm at the Python for Physics day now and I don't have the time to write
more, but I'll be glad to continue the discussion later.
On Mon, Aug 29, 2011 at 11:25:06AM +0200, StÃ¯Â¿Â½fan van der Walt wrote:
> Hi all,
> After merging the GSoC contributions, I think we're in good shape for
> a next release. Before we get there, however, I'd like to ask your
> Should we declare OpenCV a dependency? There are two sides to this:
> 1) I want to stop focusing on optimising low-level routines, and
> instead implement interesting algorithms but 2) OpenCV is quite a big
> dependency (although it does run on Windows, Linux, Mac etc, comes
> with built binaries in all of them, and has a permissive BSD license).
> Fact is, our manpower is limited, so we should focus on where we can
> make a difference. My aim is to provide an excellent image processing
> toolbox for Python that at least covers the MATLAB functionality and
> more, and to reach that goal with our limited manpower, we need to
> carefully pick our strategy.
> I'd be glad to hear what you have to say.