@VitaliyKaurov I'm pretty sure you can guess who he is: In this post he refers to a question on SE. Looking at the answers there, I suspect it is this guy.

@Kuba belated reply to your question 11 days ago. I didn't entirely follow what you wanted to do. I define DM widgets and other sub-functionality in *.m files and import them into the main DM via Initialization.

Here's an image processing challenge for anyone who is interested!! @VitaliyKaurov @halirutan @belisarius @nikie

Create high-res images better in Mathematica. The idea: shoot several pictures of a static subject in a burst, handheld. The images won't be perfectly aligned. Combine the images with Mathematica to get higher-resolution, moire-free result.

The idea is from the article I linked, which shows how to do it with Photoshop, using simple upsampling and averaging. We could do better than that. Recommended reading about how some cameras do this in a controlled and automated way.

@Szabolcs ImageCorrespondingPoints + FindGeometricTransform + ImageTransformation + ImageCompose (in linear color space)?

Or even ImageAlign...

I think Google has given some indication they would be pushing towards this getting done automatically, but of course in considerably more intelligent manner...

If you want to get really technical... cs.technion.ac.il/~elad/software first one on the list :-)

hey guys! I often want to apply a function to each of the elements of a list, so I do myfunction@#&/@mylist

is there anything cleaner than doing that @#&/@?

looks like a curse in comics too :)

@Sosi f /@ list?

D:

@blochwave @kirma It's not interesting without getting technical. What I was interested in is doing it better, not reproducing what he did in Photoshop. Thanks for the link!

it was right there before my eyes D: thanks @Kuba

:)

@Sosi You are welcome :)

@Szabolcs There are a lot books and ongoing research about super-resolution. Implementing a nice method for this requires a lot of reading in todays published work and well, I simply averaging aligned images is only a very basic solution.

Scholar search for SR surveys

@halirutan How much do you know about the technique? In the meantime I got interested in trying to do this for practical reasons too, not just to play and learn. photoacute.com is a software that seems to do a pretty good job (I tried the trial version), but it's expensive, so I think I'll just play with Mathematica or stick to plain averaging with Photoshop for now.

@Szabolcs Not much, since I never had time to implement something myself. I read a survey once since I got interested as you. I have a book at work but that's it. Some of the difficulties are very easy to understand though.

For instance is the special re-alignment of the low-res images crucial.

@Szabolcs The best thing is really to read a survey and try to guess how much work it is to implement a good method.

@Szabolcs Btw, I'm sure you know this already, but Superresolution is closely connected to image deconvolution in several aspects.