Hi Irena,

The general answer is: Yes, of course this is interesting. However, please keep in mind that many research papers might not be easily implementable into GIMP in one summer. Many computer vision papers present conceptual ideas rather than algorithms productively usable by an end user.

Therefore my suggestion would be to pick one of your suggested papers, create a paragraph or two on what you want to do and propose it to this mailing list. This way, people who are familiar with the concretely affected parts of the GIMP code can tell you what is doable and how.

Gerald

On Mon, Mar 23, 2009 at 8:55 AM, Irena Damsky wrote:

would appreciate a reaction,
anyone?

On Sun, Mar 22, 2009 at 11:02 PM, Irena Damsky wrote:

Hi all,

I'm a Msc student of CS in the Inter diceplinery center in Herzliya, Israel.
my main interests are image proccesing, image manipulation, computational photography and 2nd computer graphics.

from the list of suggested ideas for the GIMP GSoC I saw that the main project suggestions are mainly for GUI and stuff like that, are there any project idea's that involve image manipulation?

for example, there had been papers on "gradient domain editing", such as http://www.cs.tau.ac.il/~tommer/extra/adv-graphics/pie2003.pdf and I wonder, can something like this, or like "colorization by example" such as http://www.cs.tau.ac.il/~dcor/online_papers/papers/colorization05.pdf or "natural colorization" such as
http://luanqing1021.googlepages.com/research

can something like that become a GSoC project?

Thank you! Irena

--
Irena Damsky
irena.damsky@gmail.com
++972-57-8165528
++972-52-3294417
you can also find me on MSN: ira_damsky@hotmail.com

Smile! and the world will smile with you! Cry! and you'll cry alone...

--
Irena Damsky
irena.damsky@gmail.com
057-8165528
052-3294417
you can also find me on MSN: ira_damsky@hotmail.com

Smile! and the world will smile with you! Cry! and you'll cry alone...

_______________________________________________ Gimp-developer mailing list
Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer

Hello,

My initial reaction would be that if we are to make the Gimp "production ready", then a few interesting "feet on the ground" features would be interesting to work on :
- Resize
- Denoise / Sharpen
... before exporting to CMYK.

Here are some pointers : - Fractal resizing
( http://ecommons.uwinnipeg.ca/archive/00000026/ and http://search.cpan.org/~rkobes/GD-Image-CopyIFS-0.25/CopyIFS.pm)

- Wiener denoising (Refocus it) enhancement and port to GEGL : http://refocus.sourceforge.net/doc.html

Otherwise, I could suggest - Highlights / Shadows compression and other non linear contrast enhancement techniques (pointers on request) beyond Retinex.

Another suggestion would be to improve the smoothen/sharpen tool by using wavelets decomposition and making it available as a brush : http://gimp.foebud.org/node/11742

What field of image processing are you more interested in ? We can always find ideas of something which would fit a summer ...

keep smiling :)

vincent

On Mon, Mar 23, 2009 at 16:55, Irena Damsky wrote:

would appreciate a reaction,
anyone?

On Sun, Mar 22, 2009 at 11:02 PM, Irena Damsky wrote:

Hi all,

I'm a Msc student of CS in the Inter diceplinery center in Herzliya, Israel.
my main interests are image proccesing, image manipulation, computational photography and 2nd computer graphics.

from the list of suggested ideas for the GIMP GSoC I saw that the main project suggestions are mainly for GUI and stuff like that, are there any project idea's that involve image manipulation?

for example, there had been papers on "gradient domain editing", such as http://www.cs.tau.ac.il/~tommer/extra/adv-graphics/pie2003.pdfand I wonder, can something like this, or like "colorization by example" such as http://www.cs.tau.ac.il/~dcor/online_papers/papers/colorization05.pdf or "natural colorization" such as
http://luanqing1021.googlepages.com/research

can something like that become a GSoC project?

Thank you! Irena

--
Irena Damsky
irena.damsky@gmail.com
++972-57-8165528
++972-52-3294417
you can also find me on MSN: ira_damsky@hotmail.com

Smile! and the world will smile with you! Cry! and you'll cry alone...

--
Irena Damsky
irena.damsky@gmail.com
057-8165528
052-3294417
you can also find me on MSN: ira_damsky@hotmail.com

Smile! and the world will smile with you! Cry! and you'll cry alone...

Vincent:

Do you think you could put some of your ideas for useful/interesting GSoC projects here:

http://wiki.gimp.org/gimp/SummerOfCode2009ideas

?

Nicolas Robidoux Universite Laurentienne

Irena Damsky wrote:

would appreciate a reaction,
can something like this, or like "colorization by example" such as http://www.cs.tau.ac.il/~dcor/online_papers/papers/colorization05.pdf

A colourization plug-in could be interesting. Not sure if it could be done in the time frame of GSoC. There would be a certain amount of work needed just on the implementation of the UI that let you mark parts of the image that would get certain colours.

On Mon, Mar 23, 2009 at 11:43 PM, Irena Damsky wrote:

Actually this semester I'm working on a project for school that a part of it will be the implemantation of this algorithm. I thought more in the lines of implementing the "automatic" coloring algorithm that colors the image based on an example and then using user refinment enhances the result to make it "perfect". Actually the user refinment of the result picture is almost as the "manual" coloring scheme --> on a result image that was colored using the algorithm you use brush strokes to "enhance" parts that we're not colored correctly... so, in a black and white image, by using brush strokes you can spesify that a certain region should be for example green using the same functunality of the algorithm.

unfortunatly, I've almost never wrote GUI implementation, niether in my work place nor in any school work (only once in a project I had to hand in and it was written in Matlab) thus the GUI of the implemantaion will be a good lesson to me. is it possiable to submit a project proposal and in it to extand the point that I'll need more mentoring in the GUI implemantation?

I'm sorry to spoil the party, but...

http://my.opera.com/area42/blog/gimp-colorize

What could be interesting is implementation of http://people.csail.mit.edu/soonmin/photolook/

Alexandre