New Mathematical Methodology using Sparsity

suggest a way to improve Neat Image
Post Reply
Dan
Posts: 26
Joined: Fri Feb 07, 2003 6:00 am
Location: Chicago, IL

New Mathematical Methodology using Sparsity

Post by Dan »

Please read the entire article. Could you use this technique to recover poorly sampled/noisy images?

http://www.wired.com/magazine/2010/02/f ... ithm/all/1

Thanks,

Dan
NITeam
Posts: 3173
Joined: Sat Feb 01, 2003 4:43 pm
Contact:

Post by NITeam »

It is not very suitable for modern digital cameras, the way their hardware capture images. The software side would have some problems too, but if hardware is not suitable to use that method then it doesn't really matter. Not to say that the method is not good, it is certainly good but good in circumstances different than what we usually deal with in digital cameras and images they produce.

Vlad
Dan
Posts: 26
Joined: Fri Feb 07, 2003 6:00 am
Location: Chicago, IL

Post by Dan »

Thanks for you quick reply. After I followed up on the article cited I learned that the image must be acquired using very special hardware which ensures that the image is sampled randomly. For anyone interested in the entire technical details the following link will take you there.


(note page works w IE but not FFox)
http://content.digitalwell.washington.e ... ecture.htm

I did not have a chance to edit my note before the time you had already replied. It's comforting to confirm, in another way once again, that you guys at NeatImage are fully up to date and awfully smart :)

Thanks,

Dan
NITeam
Posts: 3173
Joined: Sat Feb 01, 2003 4:43 pm
Contact:

Post by NITeam »

Thank you and yes you are certainly right, the hardware has to be special to support that method. Digital cameras work somewhat differently, so we have to use other methods with their images.

Kind regards,
Vlad
Drazick
Posts: 2
Joined: Sat Mar 12, 2016 11:45 am

Post by Drazick »

This method is generalized by Dictionary Learning.

To some extent, I guess you could say that Wavelets are also Dictionary based method.

So Neat Image is related to this :-).
Post Reply