Re: [cloudster-devel] Cloudster Image sample

[Yingjie XU <jdk...@gm...>]

Hi,

I have to say, my application is not to cluster images into k groups. It is
much more like a gif compress algorithm. That's why I thought they are two
completely different applications but not the same application with
different approches.

I have considered some applications in audio. However, I found it is really
difficult to make a reasonable user interface. What's your idea?

Best regards,
Yingjie XU
On Fri, May 22, 2009 at 7:02 PM, Guillaume Seguin
<gui...@en...>wrote:

> On 22 mai 2009 à 18:04:21 (+0200), Yingjie XU wrote :
> > Hello,
> >
> > My work is different from yours.
>
> Figured that, heh. A vector of 3 doubles might look a bit less meaningful
> than
> the implementation of an algorithm build on MIT studies. Not that I don't
> trust
> your approach, but I fear 3 doubles might be a bit underscaled for
> describing
> images (at least ones larger than 1x1 px)


> > The idea is from the following address:
> >
> http://www.leet.it/home/lale/joomla/component/option,com_wrapper/Itemid,50/
> > I think it is the project name get you confused.
> >
>
> Well, what I meant was not that you were working on the same thing as
> what's
> already been worked on, I meant that images were about to be covered by a
> sample, so that it would be pretty much a waste of time to work on another
> one.
> What we meant with Stéphane the other day was finding some other kind of
> data to
> work on, like sound or whatever else.
>
> PS : please keep the list in copy so that the discussion can be tracked by
> anyone
>
> --Guillaume
>
> > On Fri, May 22, 2009 at 3:31 PM, Guillaume Seguin
> > <gui...@en...>wrote:
> >
> > > Hello Yingjie,
> > >
> > > Just saw your commits about the image sample, but err, there's already
> an
> > > image
> > > sample on the way (see the Gist.NET branch), which should work pretty
> well
> > > when
> > > it's done (I mean, which isn't straight forward vector transformation
> but
> > > aims
> > > to produce meaningful vectors), so basically you're pretty much wasting
> > > your
> > > time.
> > >
> > > On another topic, about the vector nd commit you did the other day, it
> > > might
> > > basically lead to weird problems, since objects which are considered
> equal
> > > might
> > > have different hashcode, which is basically wrong. We're most likely
> > > willing to
> > > use straight equality, not approached one.
> > >
> > > --Guillaume
> > >
> >
> >
> >
> > --
> > Yingjie XU
> > Département d'informatique
> > École Normale Supérieure
> > 45 rue d'Ulm
> > F-75230 Paris Cedex 05
> >
> > "If you received this communication by mistake,
> > please don't forward it to anyone else (it may contain
> > confidential or privileged information), please erase
> > all copies of it, including all attachments, and please
> > let the sender know it went to the wrong person.
> > Thanks."
>



-- 
Yingjie XU
Département d'informatique
École Normale Supérieure
45 rue d'Ulm
F-75230 Paris Cedex 05

"If you received this communication by mistake,
please don't forward it to anyone else (it may contain
confidential or privileged information), please erase
all copies of it, including all attachments, and please
let the sender know it went to the wrong person.
Thanks."