Video: "A computational model of aesthetics"

People always like to joke about Photoshop eventually adding a big red “Make My Photo Good” button, automatically figuring out what looks good & what adjustments are needed. Of course, researchers are working on just that sort of thing:

As someone who aspires to be creative, I have mixed feelings. The idea of rating images according to precomputed standards of beauty makes me think of the Robin Williams character in Dead Poets Society excoriating a textbook that rated poetry along two axes:

Excrement! That’s what I think of Mr. J. Evans Pritchard! We’re not laying pipe! We’re talking about poetry. How can you describe poetry like American Bandstand? “I like Byron, I give him a 42 but I can’t dance to it!”

And yet, I find I’m intrigued by the idea, wanting to run the algorithm on my images–if only, maybe, to have fun flouting it. I also have to admit that I’d like to see the images taken by certain of my family members (not you, hon) run through such algorithms–if only to crop in on the good stuff.
[Via Jerry Harris]

21 thoughts on “Video: "A computational model of aesthetics"

  1. Impressive technology. I’m sure it will be a great boon to motivational poster designers and some wedding photographers.
    But we still need a human to make art; detect subject matter and understand context, then break the aesthetic rules in a way that emphasizes those things and evokes an emotional response.

  2. This is insane. I hopes this never ever gets in any application and is silently forgotten and buryed.
    But anyway, it is impressive.

  3. Even if it’s used as a “second opinion” (It’s tough, sometimes, to judge your own photos and I would love another photographer with more experience to suggest something), I think this is valuable.

  4. It’s very scary stuff, but inevitable. I’d hope they’d at least include more advanced composition rules (or user-defined rules). Otherwise all the photos will start looking the same.
    The most exciting part was towards the end, with the image synthesis. I could see this working quite well with content-aware scaling.

  5. This reminds me of a sentence from the incredibles: ”if everybody’s special, than no one is”.
    However, I do see many uses for this, not everything we design is art and this could free up some time used for tedious tasks. Image synthesis and scaling esthetically to a specific size alone are great features.

  6. The way I could imagine this becoming useful in Photoshop is as a feature similar to the crop tool overlays in Lightroom, i.e. a tool that gives you a suggestion that you may choose to use if that is what you were going for in the first place or if it is better than your original intent, or not, if it is not.
    The retargeting part could become interesting if it were not based on salient region detection or similar, but on manual selection of objects. So the user would select the individual objects using a lasso or a brush tool and then get transform handles on each of them to arbitrarily shift them around in the image. Just like using free transform on a selection, but the surroundings would be automatically updated/retargeted to move/transform the selected object.

  7. John,
    I fell threatened!
    I lost my first work to the PC and the web. I sold law books. The company did not see it coming, they went from 60% market share to 25% in 4 years.
    Technology sure is a blessing????
    Ken in KY

  8. I’ve read a lot of photographer’s blogs in which Photoshop is panned and Photoshop professionals viewed as something close to hacks. That sounds an awful lot like the arguments against this technology. Couldn’t it be just an “Auto-Compose” feature like “Auto-focus”? If this sort of thing threatens your livelyhood, maybe you should consider pumping up your skills?

  9. The lesson on basic composition in this demonstration is more useful than the software is interesting.
    Perhaps this software could be a valuable training program in photography classes where a student could apply the program to their photo to see the effect of optimizing composition.

  10. Is an impressive software … although I am sure that professional photographers will always have something special to offer I think it will be useful in companies where is a need for god image very quickly

  11. Oh boy, where to begin…
    This will do for the great unwashed what McDonald’s did to their diet. Edible, but certainly not inspired.
    There is no “creativity algorithm”.
    The “rule of thirds” is not a rule. Algorithms are the exact antithesis of creativity.
    Creativity cannot be quantified.
    These tools are part of the trend toward visual Muzak.
    Self-pleasuring machines.
    Ugh, I really am ill now. I think I’ll go back to painting.

  12. Well…
    There is a lot of potential for good in this kind of technology, and not just for amusement. Any time you impose ‘rules’ on the subjective in an effort to get something objective, you tend to end up with a whole range of things artists now want to break.
    When I try to get creative with photography or design, I look for accepted rules and make a conscious choice about what to preserve and what to break, sometimes for aesthetics, sometimes to tell a story.
    Rules in creative pursuits, after all, are there to help us communicate in a common way. But those same rules are nothing more than empirical collections of observations, based on what humans generally ‘like’ or ‘dislike’. IMHO, they are not to be slavishly applied, but rather to be exploited at the highest level.
    And a lot of aesthetics are culturally based, so a tool like this would be useful in comparing marketing messages that span different regions, or in understanding how groups may perceive the story in an image.
    Exposing rules means exposing targets – the artist has the obligation to make choices about whether to hit, miss or ignore them.

  13. This is great stuff! I could use such a feature in PS to “auto-compose” a heap of so-so photos — and *maybe* it would give me some quick ideas on improving some of them. If so, that’s great — if not, do harm done to anyone, anywhere, anytime.
    But whoever “feels threatened” by such a piece of technology isn’t much of a creative person to begin with — so no sympathy from me to the Luddite whiners.

  14. Basically what it demonstrates is a profound lack of understanding of what makes photographs great. They apparently don’t understand that compositional principles, such as the rule of thirds, are like training wheels on a bike.
    Years ago, many professional photographers rightly dismissed PPofA contest winners for exhibiting slavish adherence to a strict set of rules (cliché’s?) governing composition. The result was the contests stifled creativity in the group’s members and a homogenization of what they called good was the result. Everything looked the same!
    But submit they did to the judges’ edicts if they wanted to win and reap the benefits of the adulations that come from the keepers of proper photographic taste who ran the organization.
    On the other side of the creativity spectrum I was a photojournalist for 15 years. The thing that drove our compositional esthetic was pretty simple. Content trumps compositional rules every time. And yet, good composition drew the reader in to the photo and told the story. So it was very important. But we threw off the training wheels of the rule of thirds and the notion of some kind specific ratio, or “balance,” that had to be present. The use of negative space, light, tone, shadow, line and form, color and clarity had to become instinctual. If you had to consciously tick off a checklist of compositional rules while shooting, you were still a beginner.
    Still, as far as I’m concerned, even in art that holds true to some degree. Content is king, even if it’s simply the feeling one gets from looking at an abstract image. Art has to grab you at a deep, profound level. Where statistical analysis will always fall short of determining what is and isn’t good composition. And that reminds me of a joke:
    An engineer, a physicist, and a statistician were moose hunting in northern Canada. After a short walk through the marshes they spotted a HUGE moose 150 metres away. The engineer raised his gun and fired at the moose. A puff of dust showed that the bullet landed 3 metres to the right of the moose. The physicist, realizing that there was a substantial breeze that the engineer did not account for, aimed to the left of the moose and fired. The bullet landed 3 metres to the left of the moose. The statistician jumped up and down and screamed “We got him! We got him!”
    May all our compositions be bulls eyes.
    [Heh–great stuff, Eric. –J.]

  15. Interesting research. Whether or not it proves to be useful in some way in the real world, it has a bunch of people thinking and talking about what makes art “art”.
    We’ve long had a similar joke on audio forums about a “make hit record” button. Obviously one-click perfection is a pipe dream, but part of the basic drive of human nature is to leave no possibility unexplored, and quite often our ideas produce surprising and groundbreaking results even when they don’t meet the stated goals. For those reasons, it doesn’t make sense to condemn or shut down this type of research.
    Whatever outcome you may fear, think it through. Even if some technology comes along that makes it incredibly easy to produce great photos, there will always be people with talent and imagination who manage to surprise us with things that no algorithm could ever come up with, because algorithms don’t think or feel, and it is in that same basic human nature that we will never stop being hungry for something new.
    It’s a bit silly to think this technology will somehow make everyone a professional photographer. A more realistic viewpoint is that it might mature to the point where it can improve the quality of everyone *else’s* photos. Tools like iPhoto already have a one-click “Enhance” button (another long-running joke of the Photoshop team that is nevertheless something that exists in reality), and it actually does a pretty decent job in most cases of improving things like contrast and color balance of the photos that are being taken by most people. I probably won’t use it myself, but I bet my mom will love it.

Leave a Reply

Your email address will not be published. Required fields are marked *