RE: [boost] VIGRA - Image processing for boost

From: Fernando Cacciola (fcacciola_at_[hidden])
Date: 2001-11-13 18:00:49

----- Original Message -----
From: Beman Dawes <bdawes_at_[hidden]>
To: <boost_at_[hidden]>; <boost_at_[hidden]>
Sent: Tuesday, November 13, 2001 7:47 PM
Subject: Re: [boost] VIGRA - Image processing for boost

> At 03:10 PM 11/13/2001, Ullrich Koethe wrote:
>
>  >as the maintainer of the VIGRA library, I'd like to submit VIGRA for
>  >inclusion into boost. VIGRA stands for '(Computer) Vision with GeneRic
>  >Algorithms'. It is the first library (AFAIK) that consequently applies
>  >the principles of the STL to the field of image processing and analysis.
>  >Please take a look at
>  >
>  >     http://kogs-www.informatik.uni-hamburg.de/~koethe/vigra/
>  >
>  >Features of VIGRA include:
>  >* fully generic design
>  >* extensive documentation
>  >* liberal license
>  >* templated image data structures, image transformations, filters etc.
>  >* import/export of many image file formats (partly by using external
>  >libraries such as libtiff and libjpeg)
>  >* support for several color spaces
>  >* image analysis functions (edge detection, connected components
>  >labeling etc.)
>  >* Python interface based on boost.python (not yet in the public
>  >distribution)
>
>  From looking at the home page it seems really interesting.  It's hard for
> a non-expert like me to easily grasp of applications it would be very
> useful for, possibly useful for, or unlikely to be useful for.
>
I downloaded and studied the library once (about 6 months ago, or so)
I haven't been able to actually use it, but I've found its design and
interface *really good*. In fact, it is the best image proc library that
I've seen (but I've seen just a few)
As a developer in the field, I've found *very easy* to follow the design of
its components and algorithms, both at code level and in the documentation.

I think that it might be difficult for people outside the image processing
field to review the library; but if it is scheduled for formal review, I
will vote 'accept' right away. It has all the right tools for image
processing with the right interface; it can only remain implementation
details and minor interface aspects due to boostification.

Fernando Cacciola
Sierra s.r.l.
fcacciola_at_[hidden]
www.gosierra.com