Hullo All,

On Thu, 14 Apr 2011 09:44:14 +1000, Terry Duell <[email protected]> wrote:

Hullo All,
There is an interesting paper here <http://www.cs.nyu.edu/~fergus/research/deblur.html> which appears to go a bit further than some towards solving the deblur problem.

...and there is some later stuff based on the above here
<http://www.di.ens.fr/willow/research/deblurring/> which also includes updated Matlab code.

I have had a lot of trouble with the code from the first reference, and that may be due to the particular Matlab functions that I have had to find for use with Octave. The Matlab code with the second reference is complete, they have included their own versions of functions that aren't available in Octave, so that is a step forward. Nevertheless I have not had any success thus far in getting a solution for any of the examples provided. The indications are that it is a very compute intensive process (possibly hours?), but if the results show some promise when stitched into a panorama, it may be worth some effort to refine the process to speed it up. Not sure if that is possible. The approach in both these references is essentially 'blind deconvolution', results from which never seem to be too brilliant to my eye. Given that the original problem posed by JohnG was the result of having overlapped sharp and blurred images, which should provide a good 'leg up' to the right answer, it is a pity that there doesn't seem (at the moment) to be a way of using this with an established approach to de-blurring.

Anyone interested in this problem who would like to help me make some/better progress?


Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups "Hugin and 
other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Reply via email to