denoising algorithms / working with Github

mael mael.primet at gmail.com
Wed Oct 27 16:43:42 EDT 2010


Hello,

I had lots of things to do lately, haven't had much time to work on
the library, but
I'll do it quickly,

here are the new things: http://github.com/maelp/scikits.image
(some new filters, and some nice zooms)

here are other projects that I work on in parallel, and I'd be happy
to have some contribution:
- image viewer: this should become the image viewer of scikits.image
in our  opinion (although it is not yet complete). Basically, it is a
modulable efficient viewer, that handles zoom properly, displays level
lines, etc and should be easily extensible. This is only  a stub yet,
and I'd be happy to merge it with your viewer, in particular have the
threaded computations to have faster computations, and add the
histrograms, normalization, etc.
It can show grey and color   int / float images, and should be pretty
fast even for large images since it only computes the strictly
required part that is needed to be viewed, and has a cache to speed-up
the next recomputations (if you move the view a little, most of the
rendered image does not change, and it only recomputes what  needs
be    )

- tutorial: http://github.com/maelp/tutorial
This should become a standard way to describe scikits.image algorithms
in a dynamic way (this is basically the Qt Webkit that has been
extended to include scikits.image algorithm)
This could be a good advantage of the library: it is often difficult
to find interactive use-case for algorithms, etc, and many times, old
algorithms aren't properly documented, or we don't know in which case
they do and don't apply. If we have a standard "documentation" in
interactive format (where people can try our algorithms, have a
mathematical presentation, upload their images etc) this will help
keep  up  with a large image library.
I made it such that it is visually attractive because I'm pretty sure
this is extremely important (have beautiful documentation )


I will add new filters to the library soon, please give  some inputs
on the  new projects and feel free to extend and/or modify anything

I haven't had time to properly document everything yet, I'll do that
soon



More information about the scikit-image mailing list