Luminous Landscape Forum
Raw & Post Processing, Printing => Digital Image Processing => Topic started by: Christopher Sanderson on July 04, 2009, 06:36:00 pm
-
Click here (http://www.cs.princeton.edu/gfx/pubs/Barnes_2009_PAR/patchmatch.mp4)
-
Click here (http://www.cs.princeton.edu/gfx/pubs/Barnes_2009_PAR/patchmatch.mp4)
Wow.. thats pretty amazing. And a little scary in terms on how easy its going to make it for anyone to do massive image alteration.
-
That looks like about CS9 to me. I want one!
-
Click here (http://www.cs.princeton.edu/gfx/pubs/Barnes_2009_PAR/patchmatch.mp4)
LOL...you'll notice that this is from Princeton, not MIT where seam carving came from. It's typical that Princeton comes around and has to clean up after MIT does the cool stuff. Of course, ya gotta love that hip background music from the Princeton video...oh wait...that was iTunes playing in the background. Sorry, the Princeton vid had no music (which sucks cause the music I was playing made the Princeton vid sound really hip!!!)
Actually this was sorta tame compared to some of the stuff I saw at MIT what Eric Chan brought me around to meet the geeks when I was in town for the EPson Print Academy last March. They're working on some pretty way cool stuff!
-
I've seen vids of retargeting before, but doing it in real time would make this usable. This is just as cool as the tech demo for a camera which took a photo with editable focus point in one shot (can't find the video).
But it raises serious ethical questions which are not limited only to photojournalism, as extensive image editing becomes quick, easy and available even to beginners.
-
yes, but although this technology is very interesting, there's no reason to expect that the initial release of this functionality will work any better than their initial releases of pano stitching or dof blending, or even a couple of releases later for HDR blending.
-
yes, but although this technology is very interesting, there's no reason to expect that the initial release of this functionality will work any better than their initial releases of pano stitching or dof blending, or even a couple of releases later for HDR blending.
That's how almost all software and advanced hardware is sold these days: get it out as soon as possible for dirt cheap, and charge the users for each minor incremental upgrade, to finally come up with a fully working product by version 3. In the end you've ended up paying for a premium product, but with the additional hassle of working with and around the bugs and lack of features of a full-fledged product.
*sigh*