Content aware resizing (using the seam carving algorithm) is also available as a plugin for GIMP:

http://liquidrescale.wikidot.com/

Here's the presentation (SIGGRAPH 2007) from the authors of the library used in the plugin:

http://www.youtube.com/watch?v=vIFCV2spKtg&feature=playe...

Heck, you can do liquid rescale in Javascript now. http://labs.pimsworld.org/wp-content/uploads/2009/04/demo-co... (requires a browser with fast javascript - not IE)

I wonder if techniques similar to this will be included somehow in CS5:

http://www.cs.washington.edu/homes/pro/papers/videoEnhanceme...

Both of these techniques are drool-worthy. As a designer who spends a lot of time using the creative suite, I want both features YESTERDAY. Nice post.

Well done. But, at what point of amazement do Adobe engineers have to worry about being burned as witches?

We implemented a seam carving algo in our computer vision course. We found "perfect" (minimum cost) seams, so the algo was very slow. It's great that more research has gone into faster (although perhaps not as theoretically "perfect") algorithms.

If we're allowed to destroy our pictures by non-ratio-holding resizes, then we should be able to destroy our pictures with cool seam carving techniques. Think of the possibilities!

seam curve is not slow. I implemented seam curving to shrink 800x600 image into 800x200 image on one core computer with less than 1 second.

Edit: if you mean the total minimum cost of all curves, I am wrong.

Yep, total minimum cost of all curves is what I did. I did it in Matlab, something I don't know much about at all, so it is definitely not optimized.

"coming in CS5" is misleading.

CS5.org is not run by Adobe, who haven't said anything about this being included in CS5 products.

Links to the researchers if you are interested:

http://www.danbgoldman.com/

http://www.wisdom.weizmann.ac.il/~elishe/

http://www.connellybarnes.com/work/

http://www.cs.princeton.edu/~af/

Great work guys.

I haven't even learned one tenth of the features in CS3 yet. And yet, surprise, surprise, I seem to do just fine. More feature bloat as far as I'm concerned.

There are features I do not use. Therefore, new features will be of no use to me.

This line of reasoning seems rather lacking.

I wish Adobe would just create a single program I buy ONCE for $499 instead of every couple years, then allow me to buy only the add-ons that I need at $40 each (for example).

I guess I can't win against corporate greed, and since they have no REAL competition after buying Macromedia, they will keep getting away with these outrageous prices for upgrades.

If you don't think the features are worth the price than don't buy the product. Adobe aren't forcing you to upgrade. If enough people don't think the features are worth buying than Adobe will reduce the price.

But frankly CS seems cheap at the price. $500 - that's probably what 5 hours of your time costs to your employer, the amount of time the new CS features will save most professional users means it'll pay for itself within weeks if not days.