Content Aware Image Resizing

An answer (at least a potential one) to the dilemma of fluid layouts and fixed ratio images in web design.

Watch the whole thing to see some really cool possibilities.

8 Comments

  1. nstryker
    Posted September 6, 2007 at 9:23 pm | Permalink

    the removing people angle at the end is kinda cool/creepy, but their initial examples of altering nature are definitely just as bad to me as squashing someone’s face up. i don’t really see a difference there, between mushing nature’s perspective vs. mushing someone’s face…especially having been to some of the locations they show.

  2. Posted September 6, 2007 at 10:39 pm | Permalink

    I thought vector graphics was in place to solve the flexibility restraints of on non-image tiled designs; probably hasn’t caught steam since IE doesn’t completely supported SVG.

    I agree with Nathan, I saw this a week or two ago on TechCrunch when he was asking for this functionality in Photoshop. I didn’t get the benefit of smashing pictures then and I still don’t, even if Adobe hired the guy.

    It would be cool for that average user who doesn’t care about integrity in their photos and/or doesn’t know how to just use the tools already provided in PS or iPhoto or whatever.

  3. Posted September 6, 2007 at 11:07 pm | Permalink

    I think the difference there is that there are lots of situations where you (the designer) might be OK with the nature shots being altered, whereas it’s almost never OK to do that with faces because they just won’t look right.

    Of course, if the intent of the photos is to give a completely accurate picture of the scenery, you wouldn’t want it to be altered, but there are probably a lot of times where that isn’t the case.

  4. nstryker
    Posted September 7, 2007 at 7:22 am | Permalink

    why would you have a photo if you didn’t want to accurately represent what the photo is of? that’s the whole point of a photo!

  5. Posted September 7, 2007 at 7:47 am | Permalink

    Depends on the intent of the artist / designer. Sometimes someone may just want a generic picture of *a* beach (as opposed to a specific one), in which case it doesn’t matter if it isn’t accurate as long as it still looks realistic, which is the point of this technique.

    Even for representing a specific place, some people may be OK with eliminating the least important parts while preserving most of what makes the image unique, for the purpose of altering the proportions of the image to better fit the layout.

    I take it you wouldn’t be interested in using this technique for photos, but I imagine lots of other people will, and that doesn’t necessarily make them “wrong”, does it?

  6. Posted September 7, 2007 at 8:14 am | Permalink

    Why am I always being moderated?

  7. Posted September 7, 2007 at 2:19 pm | Permalink

    Re: the moderation - I’m not sure, but one difference I did see between your comment that didn’t get moderated and the one that did is you used a bogus e-mail address in the first one. Looks like the same thing we talked about before.

    Vector graphics are a whole different situation; this technology is meant for pixel based images like photos, which you’re not going to use vector to display.

    For the other points, see my follow-up comments above addressing the utility of this technique for photos being used in all sorts of contexts. Sure, it’s not something you’d want on every image, but it could come in handy in some cases.

  8. Posted September 7, 2007 at 5:16 pm | Permalink

    Fancy effects have long been a part of artistic photography, but doing this with journalistic photographs is crack. Yes, very evil!

Post a Comment

Your email is never shared. Required fields are marked *

*
*