No subject

must be the most important factor here...

I will make a small test program to compare...

Gilles

------=_Part_11060_9889508.1201538509272
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

<br><br><div><span class="gmail_quote">2008/1/28, Marcel Wiesweg <<a href="mailto:marcel.wiesweg at gmx.de">marcel.wiesweg at gmx.de</a>>:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
> * New features :<br>>      - Added functions 'CImg<T>::get_haar()', 'CImg<T>::haar()' allowing to<br>> compute the direct and<br>>      inverse Haar multiscale transform on 2D or 3D images.<br>
><br>> Like you know, we will store and use Haar matrix into database to perform<br>> image fuzzy searches and duplicate items).<br>><br>> What do you think to use this way to compute Haar matrix ?<br>><br>
> Alternative way is to import Haar algorithm from ImgSeek :<br>><br>> <a href="http://imgseek.svn.sourceforge.net/viewvc/imgseek/trunk/imgseek-python/imgS">http://imgseek.svn.sourceforge.net/viewvc/imgseek/trunk/imgseek-python/imgS</a><br>
>eekLib/haar.cpp?revision=56&view=markup<br>> <a href="http://imgseek.svn.sourceforge.net/viewvc/imgseek/trunk/imgseek-python/imgS">http://imgseek.svn.sourceforge.net/viewvc/imgseek/trunk/imgseek-python/imgS</a><br>
>eekLib/haar.h?revision=56&view=markup<br>><br>> ... But this can be more complicated to do...<br>><br>> Your viewpoints ?<br><br>If there are no strong reasons not to use CImg in this case (very slow), which<br>
I dont believe - in ImgSeek it was written in Python with native code parts<br>after all - then it comes just right in time to save us some time, doesnt it?</blockquote><div><br>From ImgSeek, there is only haar algorithm witch have been written in C++ like you can see.<br>
<br>From CImg, i don't know if algorithm speed-up. And, you have right, speed must be the most important factor here...<br><br>I will make a small test program to compare...<br><br>Gilles</div></div><br>

------=_Part_11060_9889508.1201538509272--