Stories
Slash Boxes
Comments

SoylentNews is people

SoylentNews is powered by your submissions, so send in your scoop. Only 17 submissions in the queue.
posted by CoolHand on Saturday August 15 2015, @12:58AM   Printer-friendly
from the software-can-make-coffee-too dept.

Everyone may be a critic, but now Penn State researchers are paving a way for machines to get in on the act. However, the researchers add that their photo-analysis algorithm is designed to offer constructive feedback, not to replace photographers.

The researchers have developed an algorithm that analyses the arrangement of visual elements—the composition—of digital photographs. It also offers feedback about the perceived composition of the photograph and provides examples of similarly composed pictures of high aesthetic value, said James Wang, professor of information sciences and technology. Wang and colleagues recently received a patent for the system. "If you think about aesthetics, everything is about composition," said Wang. "You can look into colours, or textures, or shapes, but, if you boil it down, you eventually have to consider all of these elements as part of composition."

Training a machine to become an art critic is not easy, according to the researchers. A machine must be trained with examples of highly regarded photographs in order to recognize good compositional elements, said Jia Li, professor of statistics, who worked with Wang.

The original article can be found at Phys.org.

The original source can be found at Penn State University.


Original Submission

This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 0, Interesting) by Anonymous Coward on Saturday August 15 2015, @02:00AM

    by Anonymous Coward on Saturday August 15 2015, @02:00AM (#223109)

    Soon they will "teach" their algorithm to critique photographs from non-friendlies as unworthy. But first they have to wait for their system to be adopted and get at least some respect. Then anyone not sharing the official view will have their photographs disregarded and rejected.

    I am just not comfortable with anything making decisions automatically. Too easy to work around for someone who knows how it works, and too easy to rig. And there is almost always an evil purpose behind it.

    This is what is called "objectification of creativity".

    • (Score: 2) by Tork on Saturday August 15 2015, @02:33AM

      by Tork (3914) Subscriber Badge on Saturday August 15 2015, @02:33AM (#223120)

      This is what is called "objectification of creativity"

      There are situations where that might be true, but not in this case. In this context composition is all about making a 2D image of a 3D scene readable. Yes, there is creativity in composition, but no, 'unreadable' is not another form of creativity. This is more like testing handwriting for legibility than judging one's skill as a writer.

      --
      🏳️‍🌈 Proud Ally 🏳️‍🌈
      • (Score: 3, Interesting) by frojack on Saturday August 15 2015, @03:50AM

        by frojack (1554) on Saturday August 15 2015, @03:50AM (#223130) Journal

        I think you have read way too much into it.

        It is nothing lie what you suggest, the system has no idea of the 3D scene, and is presented only with a 2d picture. Using this picture, and a whole bunch of similar pictures that it has been told are compositionally pleasing, it puts out comments about how to make the picture more like the similar pictures.

        So all it is doing is automating the critique of pictures, similar to what any photography instructor would do with first year photograph students to pass down his interpretation of a pleasing layout. You can find examples of these composition rules [photographymad.com] all over the net. But in this case, rather than the system knowing the rules, it simply has a bunch of similar pictures to use as good (and perhaps bad) examples.

        So there is even less knowledge built into the system than you might expect. Given a bunch of instagram photos as examples, the system would probably recommend steps to change a perfectly good photo into a amateurish construction of intentionally imposed horrible filters, bad lighting, and crappy composition,

        --
        No, you are mistaken. I've always had this sig.
        • (Score: 0) by Anonymous Coward on Saturday August 15 2015, @09:04AM

          by Anonymous Coward on Saturday August 15 2015, @09:04AM (#223198)

          I'm sure all great photographers just are simply following the rules better than everyone else. Just like Shakespeare, Bach, Picasso were expertly following the rules for how to create art. I recommend reading Zen in the Art of Motorcycle Maintenance which talks about this - it's really cool book.

  • (Score: 2, Funny) by Non Sequor on Saturday August 15 2015, @03:02AM

    by Non Sequor (1005) on Saturday August 15 2015, @03:02AM (#223125) Journal

    Automated criticism is a solved problem:

    http://www.insultgenerator.org [insultgenerator.org]

    --
    Write your congressman. Tell him he sucks.
  • (Score: 2) by Snotnose on Saturday August 15 2015, @04:59AM

    by Snotnose (1623) on Saturday August 15 2015, @04:59AM (#223147)

    and it only looked at my photos, the code would boil down to:

    System.out.println("sucks");

    / Actually got to be a decent photographer some 30 years ago
    // Couldn't afford developing the pictures
    /// Slides were cheaper, but still cost too much and a pain to look at to boot

    --
    When the dust settled America realized it was saved by a porn star.
  • (Score: 2) by cafebabe on Saturday August 15 2015, @09:26AM

    by cafebabe (894) on Saturday August 15 2015, @09:26AM (#223201) Journal

    if((metadata indicates use of filter)||(JPEG quantize table indicates use of filter)) {
        print("using a filter on a picture is akin to polishing a turd.\n");
    } else if(rand()*1.0/RAND_MAX>0.000001) {
        print("your picture is awful.\n");
    } else {
        print("your work has promise.\n");
    }

    --
    1702845791×2
  • (Score: 2) by hendrikboom on Sunday August 16 2015, @05:47PM

    by hendrikboom (1125) Subscriber Badge on Sunday August 16 2015, @05:47PM (#223571) Homepage Journal

    So what happens when you combine this with Google's image synthesis program? You know, the one that filters random noise and improves its rating so as te eventuallly, if it has been trained on coffee cans, presents a random mess of coffee can parts? Will we get a random excellent composition?