>If this is the concensus (that should be established after concerting the
scikit-image guys), then let's do it!

Is stefanv on this mailing list?  I've sent him an email asking about HoGs, which I'll post when I get a response.


On 31 August 2011 12:12, Gael Varoquaux <gael.varoquaux@normalesup.org> wrote:
On Wed, Aug 31, 2011 at 12:11:37PM +0200, Olivier Grisel wrote:
> > Where is the right place for it then?  OpenCV is an option.  It has a
> > version, but its incredibly complicated and in my view obfuscated code.
> > I'm also growing less certain about the quality of that codebase as it
> > increases in size.

> OpenCV is mainly a C++ library (with python bindings). I would go for
> either scikit-image or scikit-learn.

Agreed.

> I would rather go for scikit-learn since AFAIK there is no documented
> use of HOG for pure image processing: you need a machine learning
> model on top of them to do image detection / categorization or
> clustering.

If this is the concensus (that should be established after concerting the
scikit-image guys), then let's do it!

Gael