Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
×
Input Devices Yahoo! Technology

The Algorithm That 'Sees' Beauty In Photographic Portraits 76

KentuckyFC (1144503) writes "Beauty is in the eye of the beholder. But what if the beholder is a machine? Scientists from Yahoo Labs in Barcelona have trained a machine learning algorithm to pick out beautiful photographic portraits from a collection of not-so-beautiful ones. They began with a set of 10,000 portraits that have been rated by humans and then allowed the algorithm to "learn" the difference by taking into account personal factors such as the age, sex and race of the subject as well as technical factors such as the sharpness of the image, the exposure and the contrast between the face and the background and so on. The trained algorithm was then able to reliably pick out the most beautiful portraits. Curiously, the algorithm does this by ignoring personal details such as age, sex, race, eye colour and so on and instead focuses only on technical details such as sharpness, exposure and contrast. The team say this suggests that any subject can be part of a stunning portrait regardless of their looks. It also suggests that "perfect portrait" algorithms could be built in to the next generation of cameras, rather like the smile-capturing algorithms of today."
This discussion has been archived. No new comments can be posted.

The Algorithm That 'Sees' Beauty In Photographic Portraits

Comments Filter:
  • Great.... (Score:5, Insightful)

    by funwithBSD ( 245349 ) on Thursday February 05, 2015 @11:51AM (#48989209)

    Now all pictures will tend to be the same with the algorithm telling the amateur photographer how to frame the shot.

    • The future of fine art. Ugly pictures because those are the only ones not generated by an algorithm.
      • The future of fine art. Ugly pictures because those are the only ones not generated by an algorithm.

        You've just described all the *ism movements at the beginning of the XX century. Your clock is 100 years late.

        • Well as they say "Those who cannot remember the past...". Did I hear bell bottoms are coming back into style?
      • Beauty is only skin deep....

        Ugly is to the bone....

        Beauty always fades away....

        But UGLY holds its own!!!

    • Re:Great.... (Score:5, Informative)

      by gstoddart ( 321705 ) on Thursday February 05, 2015 @01:09PM (#48990003) Homepage

      You know, if amateur photographers would learn how to frame shots, that would be great. It would make most pictures suck less. The framing has been well described for decades. But there's more to it than just that.

      But if you think all of the hacks with iPhones are suddenly going to make beautiful portrait photography, you underestimate just how bad most amateurs are.

      In a world where the selfie stick is a real thing, I'm pretty sure the pros have nothing to worry about it your average person with a cell phone magically learned how to frame a portrait.

      Indeed, the single most important factor is the sharpness of the image. But other important factors include the contrast between the face and background. Curiously, exposure quality is negatively correlated with beauty suggesting that photographers can create beautiful images by playing with under and overexposed images.

      By the time you're talking about professional (or really good) photographers doing "fine art" photography, and understanding the mechanics of cameras it's simply a different thing.

      And, I say this not as a "pro", but someone who has been taking pictures for much of his life -- taking decent pictures is more than just pointing the camera, and taking beautiful pictures involves a lot of technique.

      Even gear isn't a guarantee ... I've seen people take shit pictures with an SLR, and I've seen people take quite good shots with a point and shoot.

      • Re:Great.... (Score:5, Insightful)

        by Cederic ( 9623 ) on Thursday February 05, 2015 @02:27PM (#48991039) Journal

        When it comes to portraits, framing is piss easy. I know, even I can manage it.

        The pros earn their money because of the lighting.

      • by tlhIngan ( 30335 )

        Even gear isn't a guarantee ... I've seen people take shit pictures with an SLR, and I've seen people take quite good shots with a point and shoot.

        Generally speaking, this has always been true. A good photographer can take excellent photos with a cellphone or a disposable. While a good camera cannot turn a snapshot into a photo.

        The same rules of composition, lighting, framing, etc., apply whether or not you use a dSLR, a point and shoot, or a cellphone camera. While limitations in each may make taking some

      • by Smauler ( 915644 )

        The best pictures are almost always taken by amateurs, because the best situations are so rare, and at least 99% of photos are taken by amateurs now.

        Personally, I don't think there are many photos better than the crasher squirrel [typepad.com].

    • Now all pictures will tend to be the same with the algorithm telling the amateur photographer how to frame the shot.

      You say that as if it was a bad thing.

    • by Tablizer ( 95088 )

      Then they sell you the upgrade which can detect images ruined by the last versions' over-used visual memes. It's similar to how executives want the latest Powerpoint to not look all the other Powerpoints. Gatesian brilliance.

      • A reworked version of last sentence:

        "It's similar to how executives want the latest Powerpoint version in order to not look like all the other Powerpoint presentations floating around."

    • Now all pictures will tend to be the same with the algorithm telling the amateur photographer how to frame the shot.

      Sorta like the way most movies are all orange and blue now, huh?

    • Now all pictures will tend to be the same with the algorithm telling the amateur photographer how to frame the shot.

      What makes you think this didn't already happen 20 years ago?

      Really go into any "professional" studio and you're likely to find the same thing over and over again. Umbrella 45deg left, umbrella 45deg right, 60/40 split between the lights, both elevated, a small flash for a hair-light for background separation, the camera set to f/11, have the subject look at the camera but not face it flat on, sitting with hands in lap, press the button and nuke it till it glows.

      The algorithm is used everywhere. You'll find

  • by hcs_$reboot ( 1536101 ) on Thursday February 05, 2015 @11:52AM (#48989225)
    From TFA

    The good news for many of us is that that the visual qualities of the person in the photographic have little impact on the beauty of the resulting portrait. “We find that race, gender, and age are largely uncorrelated with photographic beauty,” Instead, many of the factors that do correlate with beauty relate to the technical quality of the image. “Aesthetic score is related to sharpness of facial landmarks, image contrast, exposure, homogeneity, illumination pattern, uniqueness, and originality,”

    So, no worries, here.

  • How would the algorithms deal with that?

    https://www.google.com/search?... [google.com]

    • I'm not quite sure why, but I think for lunch I'll have an entree of potato soup followed by boiled potatoes, mashed potatoes, patatas bravas,
      roast potatoes, gnocchi, potato skins, latkes, hash browns, hasselback potatoes, home fries, oven fries, potatoes au gratin and baked potato with sides of tater tots, french fries and potato salad, finishing with potato chips for dessert.

      Source: http://www.huffingtonpost.com/... [huffingtonpost.com]

    • Marilyn Monroe could pull off a potato sack

      She could, but for some reason she's kept it on in all of those photos.

  • by nimbius ( 983462 ) on Thursday February 05, 2015 @11:54AM (#48989253) Homepage
    We should be teaching machines to help us, not emulate our flaws and biases. beauty and ugliness, rich and poor, black and white, these are all concepts that in technology hold no relevance or meaning because theyre metrics by which we categorize and qualify people in often arbitrary and flawed ways.
    • Re: (Score:2, Informative)

      by oodaloop ( 1229816 )

      beauty and ugliness, rich and poor, black and white, these are all concepts that in technology hold no relevance or meaning because theyre metrics by which we categorize and qualify people in often arbitrary and flawed ways.

      And yet men the world over prefer women with waist/hip ratios between .68 and .71. What we consider physically attractive is directly tied to health and fitness to bear/raise children, not some arbitrary value.

      • Re: (Score:3, Insightful)

        Except during the 1920s, when the ratio was closer to 1, and during Rueben's time when, well... and actually are you sure about that "the world over" thing? Or even in just America? I mean, look at different cultures within the US.

        Definitions of physical beauty are constantly changing, and seem to be inherently based upon social, not biological, processes.

        • The ratio holds for all body types, skinny or fat.

          http://en.wikipedia.org/wiki/Waist%E2%80%93hip_ratio
          • The 1920s in particular wasn't just about thinness but about lacking curves altogether. Even breasts, which is arguably the only "men look at them and think 'Great potential mother'" attribute mentioned, were deprecated and supposed to be flattened and hidden. The hip:waist ratio really was close on 1:1 during that period.

        • Except during the 1920s, when the ratio was closer to 1 .......Definitions of physical beauty are constantly changing, and seem to be inherently based upon social, not biological, processes.

          You are assuming that men liked women looking like they did in the 1920's. From what I have heard they did not. It has always been women who choose mainstream fashion trends among themselves, not men. Look at the differnce between the fashion seen in porn (at the start before they take their clothes off) and fashion seen generally around.

          There was one period in modern times when mainstream fashion aligned with what men liked - around 1970 with the Miniskirt and fishnets. When fashion moved on to the

    • We should be teaching machines to help us, not emulate our flaws and biases.

      To help us, they must know us.

    • We should be teaching machines to help us, not emulate our flaws and biases.

      What forces us to choose?

      We should be teaching machines everything we can. They might be our only way to conquer the universe.

      We should explore all routes, every corner, everything. The only useless knowledge is the perishable one, like how many atoms are in this precise sea wave. Everything else, we shall know.

      • If we don't know how many atoms are in this precise sea wave, we won't be able to conquer the universe. How else are we going to manipulate time and space so we can move off this rock if we don't know such things?

    • Re:a wasted effort. (Score:4, Informative)

      by bws111 ( 1216812 ) on Thursday February 05, 2015 @12:28PM (#48989611)

      Did you actually read all the way through the summary? For that matter, did you even read the title? It finds beautiful PORTRAITS, not beautiful SUBJECTS. It does this by IGNORING age, sex, etc and concentrates on sharpness, contrast, and exposure. What exactly is the problem?

  • by alen ( 225700 ) on Thursday February 05, 2015 @11:59AM (#48989287)

    i know some asian people and their idea of beauty is different than european. but for us europeans, this is old news. most models used on magazine covers and beauty queens have had similar facial shapes and features for decades. i remember seeing feminists on Maury and other talk shows in the 80's complaining how the media had a narrow view of beauty.

    • by bws111 ( 1216812 )

      Neither the summary nor the article say anything about determining the 'beauty' of the SUBJECT. They are only determining the beauty of the PICTURE, by looking at things like sharpness and contrast.

    • by swb ( 14022 )

      I seem to a recall a documentary or article discussing the biological drive in reproduction that referred to a study that did a cross-cultural analysis of physical features that each culture valued. IIRC, they found that there were certain physical traits like waist-hip ratio that were independent of any culture preferences.

      OT: When my wife were in Fiji 15 years ago, we took what amounted to a water version of a bus (stopped at many islands) in order to get to a small resort island we were staying at for

  • I've seen the classifications of other algorithms and they fall prey to the distortions of makeup. For example an average looking woman with makeup accentuating features, particularly dark eyebrows and eyelashes, will score higher than a beautiful woman without. The fact that the score can be affected by image quality suggests it's using the same sort of NN features. Better features for beauty are symmetry, facial width to height ratio, unblemished skin, bone structure, etc. ANNs tend not to understand mo
  • Proof that photos and real life can and do differ?

    Hmmmm.... Imagine that. Congratulations on rediscovering what art discovered centuries ago.

    This idea that the technical details of an image have a large affect in how we perceive the subject of the photo has been well understood for decades, even if it's not be well practiced by your every day "point and shoot" photographer. Just walk into even a low end portrait studio and think about how they all use the same kinds of lighting, flash, backgrounds, foca

    • You know, by the same token I've seen some of the "fine art" wankers saying outright "Photography can never be art because it's merely an objective representation taken by a machine".

      And that reminds me why the "fine art" crowd can be as ignorant and clueless as anybody else by not understanding the medium. There's more to it than the pure mechanics.

      A can of "Merde de L'Artist" (shit of the artist) can be art, but a photograph can't? Sorry, but don't expect us to take you seriously. with that claim. It m

    • They didn't rediscover it, they created predictive algorithms by leveraging machine learning. That's a pretty cool accomplishment similar to auto-focus and face detection algorithms.

      It's entirely possible that comparing their results to traditional photograph and compositional "rules" may result in even better rules than the time-trusted ones we use now.

  • The trained algorithm was then able to reliably pick out the most beautiful portraits. Curiously, the algorithm does this by ignoring personal details such as age, sex, race, eye colour and so on and instead focuses only on technical details such as sharpness, exposure and contrast.

    In other words, the computer was unable to learn to distinguish beautiful people from ugly people.

    • In other words, the computer was unable to learn to distinguish beautiful people from ugly people.

      Why should it do that? It was not the purpose of the experiment.

  • ... and instead focuses only on technical details such as sharpness, exposure and contrast. [ ... ] It also suggests that "perfect portrait" algorithms could be built in to the next generation of cameras, rather like the smile-capturing algorithms of today."

    Algorithms that result in a properly exposed and in focus picture? I think that's been done.

  • [...] then allowed the algorithm to "learn" the difference by taking into account personal factors such as the age, sex and race of the subject as well as technical factors such as the sharpness of the image, the exposure and the contrast between the face and the background and so on

    Curiously, the algorithm does this by ignoring personal details such as age, sex, race, eye colour and so on and instead focuses only on technical details such as sharpness, exposure and contrast.

    Something does not compute

    • by bws111 ( 1216812 )

      Why? They asked people to rate the images based on personal factors AND technical factors, and found you could get the same ratings using only technical factors. What is wrong with that?

    • Serves me right for reading only the summary before commenting. I knew /. summaries were bad but this is just dumb. The article does not say that the algorithm "Curiously, ignores personal details such as age, sex, race, eye colour and so on" at all and nor does the paper synopsis. This study says nothing about "beauty" at all... it's basically a study about photographic or artistic composition and what compositions are pleasing to the eye.

  • Scene from the film L.A. Story [wikipedia.org] in which Harris K. Telemacher (Steve Martin) describes an off-screen painting in a museum:

    • I like the relationships. Each character has his own story.
    • The puppy is a bit too much, but you have to overlook that.
    • The way he's holding her, it's almost... filthy.
    • He's about to kiss her and she's pulling away...
    • The way his leg is smashed up against her...
    • Look how he's painted the blouse, sort of translucent, you can make out her breast, and it's sort of touching him...
    • It's really pretty torrid, don't you think?
    • And of course you have the onlookers peeking out like they're all shocked.
    • They wish.
    • I must admit, when I see a painting like this, I get emotionally...
    • Erect.

    The camera turns to reveal a large oil painting of a solid red rectangle.

  • ... is that by and large, there are some technical features that most humans - or at least most humans that the scientists from Yahoo Labs in Barcelona used to train this computer - agree create a beautiful image.

  • Related: see also apps that detect whether a photo has been photoshopped. https://itunes.apple.com/us/ap... [apple.com]

  • So photographers who aren't subscribers of the f/64 school of photography (to critically summarized: technical skill of taking and making the photo is what transforms a good photograph into a great photograph) are losers. I'll keep Nan Goldin, Cindy Sherman, Henri Cartier-Bresson, Garry Winogrand, Robert Frank, Man Ray, and thousands of other photographers who make excellent works of art but capturing the decisive moment (H.C.B. and Winogrand) or obtain a level of intimacy with and about their subjects (Gol

    • ... who make excellent works of art but capturing the decisive moment ...

      ... by capturing the decisive moment ...

      Oops.

  • This type of analysis would be better for taking pictures rather than analyzing existing ones, and even more so for outdoor photography (landscapes, natural events, etc.).

    The best landscape/storm photos are about timing, you have to be there when the awesomeness happens and snap the shot.
    Think the Grand Canyon full of clouds:
    http://i2.cdn.turner.com/cnn/d... [turner.com]

    Imagine a scenario where you setup a camera at a great view/vista (somehow securely) and leave it there for a while, letting a system decide when to tak

    • by Cederic ( 9623 )

      The best landscape/storm photos are about timing, you have to be there when the awesomeness happens and snap the shot.

      Problem is, the photo you linked is from someone taking a snap.

      Sure, if you're familiar with that vista when there aren't clouds then it conveys how unusual the scene is, but as a photograph it's nothing special at all.

      A photograph like http://www.terragalleria.com/p... [terragalleria.com] would be a better example. That doesn't happen because someone with a camera just went "oh, that's cool" *click*

      • I was thinking of ways to capture the rarer moments without having to be present. The Grand Canyon full of clouds is pretty rare (I've seen much better photos than the one I posted).

        The system would also have to have a configurable way to manage exposure settings when taking a batch of photos. This is accomplished via bracketing features (which would also lend itself to HDR photos).

        There's a particular view I have taken thousands of photos at over the years, it's that type of location I'm thinking about.

        T

  • To match human responses? Simple:

    If viewer is male
      give high ratings to pictures with dogs playing poker or pool
    else
      give high ratings to pictures with flowers, sunsets, and rainbows
    end

  • The thing is that people who take the most carefully lit, composed and focussed images (which is what the computer is using as it's metric) are professional photographers. They use models who are generally considered "beautiful". So unless you're very careful about your initial data set, you're going to come to some very bad conclusions...and that's what seems to have happened here.

    So, this is bogus science...badly done.

        -- Steve

  • As anyone who has experienced Dorothea Lange's "Migrant Mother" [loc.gov] can attest, while the subject has never been described as classically "beautiful," the photograph is normally described as beautiful (and powerful) by almost everyone who has seen it.

  • by JMZero ( 449047 ) on Thursday February 05, 2015 @01:58PM (#48990595) Homepage

    My guess is what they've really determined is that:

    1. Better photographers take better pictures, and also are more competent technically (ie. they take sharp, well-lit pictures)
    2. People put more effort into getting technicals right when they're shooting something beautiful

    Taking sharper photos of dull objects will only get you so far; the correlation is due to stuff that's deeper and harder to control: the subject and the photographer's skill/effort.

  • by grumpyman ( 849537 ) on Thursday February 05, 2015 @05:15PM (#48992821)
    If (RND(1)>0.5),"beautiful", "not beautiful";

Today is the first day of the rest of your lossage.

Working...