[SciPy-Dev] question of scope wavelet code

[Luke Pfister]

On Mon, May 11, 2015 at 9:21 AM, Jon Moody <jbm139 at gmail.com> wrote:

>
>
> On 5/10/2015 4:07 PM, Ralf Gommers wrote:
> >
> > - curvelets (https://en.wikipedia.org/wiki/Curvelet)
> > - contourlets (http://en.wikipedia.org/wiki/Contourlet)
> > - shearlets (http://en.wikipedia.org/wiki/Shearlet http://shearlab.org
> > GPL)
> > - chirplets (http://en.wikipedia.org/wiki/Chirplet_transform
> > http://tfd.sourceforge.net/ GPL)
> > - ridgelets
> > (http://rsta.royalsocietypublishing.org/content/357/1760/2495.short)
> > - wedgelets (http://projecteuclid.org/euclid.aos/1018031261)
> > - noiselets (http://en.wikipedia.org/wiki/Noiselet)
> > - surfacelets (http://dx.doi.org/10.1109/TIP.2007.891785)
> > - bandelets (http://en.wikipedia.org/wiki/Bandelet)
> >
> > At least the first four items in this list seem to be quite widely
> > used. I'd like some opinions on which (if any) of these people think
> > would fit in scipy, and which ones should rather be kept in a
> > wavelets-specific package.
> >
> > Also, if anyone has experience using the contourlet toolbox under
> > discussion in issue 58, would be great to hear how it performed.
>
> It's been a while (~7 years ago), but I used the contourlet toolbox for
> denoising in PET image reconstruction.  It seemed to work better than
> regular 2D DWT but if I remember right there was an improved version of
> the contourlet (ref. 10 on the contourlet wikipedia page) that wasn't
> ever incorporated into the original matlab toolbox. Another transform I
> ended up using more was the dual-tree complex WT
> (http://en.wikipedia.org/wiki/Complex_wavelet_transform) and matlab code
> from Nick Kingsbury's web site
> (http://www-sigproc.eng.cam.ac.uk/Main/NGK). This WT has approximate
> shift invariance
> (http://en.wikipedia.org/wiki/Complex_wavelet_transform#cite_ref-4)
> which is nice when working with images.
> --
> Jon
>
> _______________________________________________
> SciPy-Dev mailing list
> SciPy-Dev at scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-dev
>

I originally proposed the inclusion of contourlets into PyWavelets,
primarily because (1) I need to use them for a project, and (2) I see the
toolbox author frequently.

I agree with Ralf's ordering.  The first three are by far the most common
in my experience with regularized image reconstruction.

I don't have much of an opinion on where the various *-let implementations
should go; only a desire to see them be ported to Python.  The lack of
advanced 2D multiscale transformations is one of the major reasons that
some of my colleagues haven't made the jump from Matlab to Python.

Luke
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scipy-dev/attachments/20150512/e6dac2d9/attachment.html>