[IPOL discuss] Démo de la méthode K_SVD

[Miguel Colom]

Dear Guedira,
I'm copying this message to the Discuss and Tech lists in IPOL, since  
I think it's worth that everybody with similar problems can see the  
discussion.

> Je vais intégrer la méthode K-SVD dans mon travail.  En fait, le
> Démo (http://demo.ipol.im/demo/45/),
> que vous avez introduit, ne marche pas convenablement. Il m'affiche  
> la plupart
> du temps des "erreurs".

(This is the Ponomarenko demo)
According to the logs of the server, there isn't any problem with the  
execution of the binary, and therefore it's likely that it's the demo  
system that is stopping the process.

I've seen your previous message about the K-SVD demo and it seems that  
it's exactly the same issue.

Normally the demo system stops the execution for two reasons:
- The demo execution takes too much time to complete. In general, no  
more than one minute is allowed.
- The program demands too much memory.

In the case of the K-SVD demo, I'm quite sure it's related to the  
memory usage policy. And also in the case of the Ponomarenko demo.

Therefore, I forward this question to the @Tech team (and also the  
Discuss list): should we allow the programs to use more resources? 1Gb  
is clearly too small for Noise Estimation and Denoising demos, even  
for small images.

Could you send me one of the images for which the demo fails?

> Ainsi, le code zippé ne fonctionne pas sur
> Windows 7. Veuillez-vous m'aider car j'ai besoin fortement des  
> résultats  pour
> mon mémoire de master. 

In this case, it seems that it's related to the compiler  
configuration, since the source code is really standard (ANSI C).

Make sure it links with the PNG and OpenMP libraries.
In any case, could you give more details about the problem with Windows 7?

Best,
Miguel