I came across this intriguing approach to image resizing via TechCrunch and The Guardian. Interesting in its own right, but it also struck me that it could be very applicable to producing schematic maps. The 4 minute video is well worth watching.
Someone else has produced an online demo based on the idea, and the paper mentioned is available at the site of co-author Ariel Shamir.
6 comments:
Wow!
It gets better than just mere demo ... rsizr.com is a free webservice to upload a photos and preform this intelligent image resizing.
really neat stuff though you have to careful about masking the areas to keep/delete in order to get good results.
It is totally brilliant, isn't it? We've been buzzing about the algorithm here at work for a little while now. So inexplicable before they explain it, so simple after you know how it's done. Awesome.
And if you liked that, you'll love the image processing magic in the Seadragon demos. This one automatically spatially locates images obtained by searching Flickr for "Notre Dame". It assembles the images into a single consistent 3D model, suspending each image in mid-air within a virtual space, such that, when viewed from any point, the cloud of surrounding images form a montage of the surrounding environment. Albeit a montage infused with delightfully surreal touches from a million different holiday snaps. Absolutely brilliant.
oooh, and the link:
http://www.ted.com/index.php/talks/view/id/129
thank you guys for the links...
so just like 3 years ago when seeing Google Earth the first time i wonder why I do research in GIS if they make this creazy things.
but appearently, somebody needs to create the image tags (used for the Notre Dame presentation).. so Geographic Information Retrieval (GIR) is on the move. ;)
I haven't checked up on your blog in awhile -- but coming here and seeing others have been all over the Seam Carving technology was pretty neat. I posted a compilation and a couple ideas I had for use over the my site this morning as well.
Post a Comment