Slashdot Mirror


Software Combines Thousands of Online Images Into One That Represents Them All

Zothecula writes If you're trying to find out what the common features of tabby cats are, a Google image search will likely yield more results than you'd ever have the time or inclination to look over. New software created at the University of California, Berkeley, however, is designed to make such quests considerably easier. Known as AverageExplorer, it searches out thousands of images of a given subject, then amalgamates them into one composite "average" image.

66 comments

  1. First post by KamikazeSquid · · Score: 3, Insightful

    Can we use it to create an amalgamation of the "average" first post on a /. article?

    1. Re:First post by NotInHere · · Score: 0

      In soviet russia, joke averages you!

    2. Re:First post by alphatel · · Score: 1

      Can we use it to create an amalgamation of the "average" first post on a /. article?

      The Average Image for your first post would look something like this:
      >Frist Post (Score: -1)
      ...

      --
      When the foot seeks the place of the head, the line is crossed. Know your place. Keep your place. Be a shoe.
  2. My fear by Anonymous Coward · · Score: 2, Insightful

    I dread that the average of all internet photos will be a pornographic picture of a woman with a penis.

    1. Re:My fear by Anonymous Coward · · Score: 0
  3. DHS by Anonymous Coward · · Score: 0

    I'm sorry to say DHS will be using this technology to determine the "archetypal average terrorist" then use the result for facial recognition. Success! Millions of hits!

    At least they can also round up all the evil cats (hint: all of them) with a similar scheme

  4. Scientific justification by sinij · · Score: 4, Funny

    So this is how a bunch of scientists justified browsing for pron and cat videos all day long? Yes, we heard this before - they are working on average algorithm for images.

    1. Re:Scientific justification by Anonymous Coward · · Score: 0

      That is compressed browsing for pron and cat videos all day long, you insensitive clod!

  5. This would be amazing for.., by Anonymous Coward · · Score: 0

    pictures of vaginas. Just saying.

    1. Re:This would be amazing for.., by ArcadeMan · · Score: 1
    2. Re:This would be amazing for.., by PopeRatzo · · Score: 0

      I've been burned by a goatse link too many times to click on that.

      Which by the way, turns out to be what you get when you combine every image on the internet together. Weird, huh?

      --
      You are welcome on my lawn.
  6. Oh goodie by smooth+wombat · · Score: 1

    I can see the searches now:

    "What does a black person look like?"

    "What does an Asian person look like?"

    "What does someone from Mississippi look like?"

    --
    We will bankrupt ourselves in the vain search for absolute security. -- Dwight D. Eisenhower
    1. Re:Oh goodie by Anonymous Coward · · Score: 0

      Does he look, like a bitch?

    2. Re:Oh goodie by PopeRatzo · · Score: 1

      You forgot, "What does a man bending over and spreading his butthole look like?"

      --
      You are welcome on my lawn.
    3. Re:Oh goodie by Tablizer · · Score: 1

      No, there's nothing average about goat-se

    4. Re: Oh goodie by Anonymous Coward · · Score: 0

      What?

  7. You'll be impressed by i+kan+reed · · Score: 3, Insightful

    With this complex algorithm that takes a fuck-ton of image data and produces for you: something that is almost impossible to tell apart from applying the blur filter on the original image.

    1. Re:You'll be impressed by turp182 · · Score: 1

      I was also going to point that out. Any graphics program can blur and image with very similar results.

      I could see a benefit to this for pattern recognition, such as determining people's ancestral makeup or what breeds a particular dog is composed of.

      The key would be well defined inputs. A large sample of each possible output value would be needed, along with details about a particular value. This would be the training (200 Labradors, 200 Beagles, etc.).

      But the next step, testing/usage, requires different software (as far as I know), but their algorithms could probably be re-purposed for it. It would take the new input and compare it against a library of averaged specific samples, somehow determining a % match for each. Then it could provide an estimate of a multi-source makeup (mixed dogs, mixed humans, etc.).

      Something of this nature, might be able to identify mixed genetics visually, or to help identify minute genetic differences in a given population (without a ton of tedious study/measurements). I'm thinking of the finches in the Galapagos Islands, that sort of thing. The "purity" of the inputs would be critical. Multiple angles would probably be necessary for animals.

      Easy to see negative eugenics type uses (and I think useful applications would be of limited value, maybe).

      --
      BlameBillCosby.com
  8. Yay! Average! by Tree131 · · Score: 3, Funny

    Suddently there's a surge in searches for "Average Penis"

    1. Re:Yay! Average! by Anonymous Coward · · Score: 0

      And up comes a photo of Tree131!

    2. Re:Yay! Average! by kwbauer · · Score: 1

      At least average is better than the half that are undersized.

  9. One image... by Deadstick · · Score: 1

    ...to rule them all?

    1. Re:One image... by Andurian · · Score: 1

      I would've gone with "One png to replace them all."

    2. Re:One image... by plover · · Score: 1

      Came for this:

      One .png to rule them all
      one wavelet to find them
      one .jpeg to compress them all
      and with the imagemagick library bind them.

      Left satisfied.

      --
      John
    3. Re:One image... by Deadstick · · Score: 1

      OK, I know when I'm outclassed...

  10. lacking detail... by Anonymous Coward · · Score: 0

    "I" could have told you comparing all internet cat pics gives you a picture of a nose that looks like mons pubis...!
     

  11. Actual article by Anonymous Coward · · Score: 5, Informative

    Here's a link to the actual article, rather than the useless link provided:

    http://newscenter.berkeley.edu/2014/08/14/average-image-for-big-visual-data/

    The video was pretty interesting!

  12. The Average Cat by sfsp · · Score: 3, Insightful

    So...what the software demonstrates is that if you line up all the pictures of cats by centering them on their noses, you will CLEARLY see...

    ...that the average cat has a nose.

    The rest is blurry and remarkably uninformative.

    There needs to be a LOT more intelligence, either machine or human, applied to this before it is remarkable.

    1. Re:The Average Cat by The+Raven · · Score: 4, Informative

      You may have read the article (dubitable), but you didn't watch the video or read the SIGGRAPH paper. They demonstrate a browsing tool that enables you to, for example, find an average nose nearly instantly. You can then filter the thousands or millions of images to find specific cat breeds, poses, situations, or colors in seconds.

      The tool is called average explorer, and it allows a user to interactively explore a vast set of image data quickly and efficiently. The one picture you describe was a single click in the explorer.

      You did the equivalent of saying "Wow, I can make a black dot on a white canvas. That's not very exciting." when presented a single click with a single tool in Photoshop.

      --
      "I will trust Google to 'do no evil' until the founders no longer run it." Hello Alphabet.
    2. Re:The Average Cat by timeOday · · Score: 1

      How does it compare to eigenfaces?

    3. Re:The Average Cat by Kjella · · Score: 1

      Somehow, I still fail to see the point... I can search for "cat" in Google Images, then if I'm not happy then "siamese cat" and finally "siamese cat jumping" because I'm probably looking for one useful picture, not a blurred mess as I'd expect trying to average what a "jump" looks like. And if you ask what an average face looks like, they mean the average feature size and location not a mathematical average. I'm trying to think of one single purpose where the results of this "average browser" is what I'm looking for and I'm coming up blank.

      --
      Live today, because you never know what tomorrow brings
    4. Re:The Average Cat by Alsee · · Score: 1

      a black dot on a white canvas

      $360,000

      -

      --
      - - You can't take something off the Internet! That's like trying to take pee out of a swimming pool.
    5. Re:The Average Cat by sfsp · · Score: 1

      Well, I did read the article. I did not immediately watch the video, and now that I have, I'm still not impressed.

      The strength of the tool is NOT the averaging of multitudes of shapes, which is what is essentially advertised. Instead, it is in finding images in the set that conform to what the user selects: filtering, not combining.

      So, the "average" of blue butterfly wings with this shape is that they are blue and have this shape. You're not AVERAGING, you're FILTERING.

      Or, given this "average" nose, find the "average" ears.

      This tool is not as demonstrated primarily an averaging tool, but a filtering tool to eliminate everything that is not arbitrarily close to the arbitrary average. I'm sure there are cases where that is useful, but it's NOT the described function.

      Automatically correlating equivalency points is nice, but not new. Morphing between the images is fun, but not new. Autoalignment of equivalency points is nice, but not new.

      Putting it all in one tool is good, though.

    6. Re:The Average Cat by Wandering+Idiot · · Score: 1

      So, it's basically a recreation of what Google's already using for their Reverse Image Search with the parameters set a bit looser?

  13. Automatically means no control by techno-vampire · · Score: 3, Insightful

    If this software searches out all images of a subject and averages them automatically, that means that there's no human control over which images to use and which to reject. Imagine what would happen if you were to let this program loose to create an average image of Shirley Temple. She started in films at the age of three and reached the age of 85, and the software would create an "average image" by mixing images of her as a small child with ones of her as an elderly woman. Even worse, there's a non-alcoholic cocktail named after her, and pictures of it would almost certainly get included.

    --
    Good, inexpensive web hosting
    1. Re:Automatically means no control by ArcadeMan · · Score: 1

      I think the result of such automated image merging will probably result in something like this.

    2. Re:Automatically means no control by Triklyn · · Score: 1, Funny

      best Ima-drink-agen EVER. combines the vivacity of youth, the wisdom decades, the trials and tribulations, loves and losses of an entire lifetime, with a splash of grenadine. Would definitely drink that in, as it were.

      i'd be more afraid of image searches for rick santorum or prince albert.

    3. Re:Automatically means no control by Anonymous Coward · · Score: 0

      The alignment is automatic. The process has a significant number of user controls. Just watch the video.

    4. Re:Automatically means no control by hythlodayr · · Score: 1

      Not necessarily. Machine learning algorithms like K-means clustering are designed for exactly this sort of problem. In principle, it can figure out that there are two different shirley temple images: Shirley Temple the human; and shirley temple the bright-red drink. Of course, depending on *how* you wash the image data into something that can be analyzed, it could make unexpected categories like "Shirley Temple black & white" vs "Shirley Temple mostly red" vs "everything else".

  14. What you combine all online images... by TheNastyInThePasty · · Score: 1

    Hint: It's porn.

    --
    The best thing about UDP jokes is I don't care if you get them or not
    1. Re:What you combine all online images... by Anonymous Coward · · Score: 0

      Everything's porn if you're determined enough.

    2. Re:What you combine all online images... by Anonymous Coward · · Score: 0

      And cats!

    3. Re:What you combine all online images... by Anonymous Coward · · Score: 0

      *blinks*

  15. Amazing by Anonymous Coward · · Score: 0

    I'm gonna try with hottest Brazilian horse cock tranny just for fun...

  16. So, where is it? by ArcadeMan · · Score: 1

    It's not online for people to play with? I wanted to see if searching for "penis" will result in screen captures of Spore.

  17. Not just you're average goat... you see? by Anonymous Coward · · Score: 0

    What I did there. Sup nurga's this website is toast :(

    1. Re:Not just you're average goat... you see? by Anonymous Coward · · Score: 0

      Maybe you should educate yourself a little.

  18. Re:Frist psot by Zero__Kelvin · · Score: 4, Funny

    Yes. That looks about correct.

    --
    Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
  19. First post by Anonymous Coward · · Score: 0

    Yeah. We have that already.

    dickbutt.jpg

  20. What did you expect? by Opportunist · · Score: 1

    Great. Just what we needed. A program that will help us find the common features of the average pussy.

    --
    We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
  21. Selfie engine. by geekmux · · Score: 1

    "Known as AverageExplorer, it searches out thousands of images of a given subject, then amalgamates them into one composite "average" image...

    Suddenly the Kim Kardashian selfie book makes total sense.

    She's not a narcissist, she's just helping test the beta.

  22. it works! by nblender · · Score: 4, Funny

    I typed in "douchebag" and it showed me a picture of some guy driving an Audi A4.

    1. Re:it works! by turp182 · · Score: 1

      True douchebaggery doesn't start until the Audi A5 in my opinion. One should also consider BMW 5 or higher and anything Mercedes in the 500/600 classes.

      --
      BlameBillCosby.com
    2. Re:it works! by Tablizer · · Score: 1

      Odd, I typed in "current location of nblender's girlfriend" and got the same thing.

    3. Re:it works! by nblender · · Score: 1

      Hope my wife doesn't type that in.

  23. dissapointed by PopeRatzo · · Score: 1

    How is it that on a story like this I'm the only one making goatse jokes?

    What happened to you, Slashdot?

    --
    You are welcome on my lawn.
  24. They must be smarter than... by Anonymous Coward · · Score: 0

    They must be smarter than the average bear.

  25. I tried it by Solandri · · Score: 2

    Here's what I got when I gave it every pic in my photo library.

    But seriously, I've seen the same technique used to discredit a movie of a UFO shot on 8 mm film. If you just watch the movie, you see an elliptical blob flying. Someone scanned the blob from each frame, aligned them, and averaged them. The increased contrast (bit-depth and resolution basically) let you see that the elliptical blob was more a diagonal prism, and that there were dark features underneath it. Basically it was a Cessna with the sun reflecting off the top of the wing.

  26. Fantastic by kuzb · · Score: 1

    They've created an algorithm for producing fuzzy blobs!

    --
    BeauHD. Worst editor since kdawson.
    1. Re:Fantastic by Anonymous Coward · · Score: 0

      This might come in handy for all those fresco restoration projects.

  27. Three Dee by Tablizer · · Score: 2

    I'd like to see one that constructs a 3D model. Perhaps it could use a genetic algorithm (GA) to breed a 3D model that can best represent the most actual specimens of the target object type.

    It may be a lot of computations, however, because one is not just running genetic algorithms, but also rotating all the candidate 3D models and lighting conditions to see which best fits the actual specimen images PER GA candidate PER specimen. Perhaps a 3D thumbnail version can be used to for initial placement estimations to be fine-tuned with a fuller model.

    Then you got spot and texture variations within specimens. You have to model varying textures, not average them out. But even if it ignores texture & spots to simplify things, a 3D shape model result would be cool.

  28. Re:Frist psot by Anonymous Coward · · Score: 0, Insightful

    Only if "you spin me right round" is playing.

  29. gizmag by Anonymous Coward · · Score: 0

    http://www.gizmag.com/averagee...

    Is it just me or does gizmag look too much like jizzmag for comfort?

  30. Interesting tidbit from the article .. by Anonymous Coward · · Score: 0

    "The authors noted that since photography was invented, there have been an estimated 3.5 trillion photos taken, including 10 percent within the past year. Facebook reports 6 billion photo uploads per month on its site, and YouTube gets 72 hours of video uploaded every minute.

  31. Re:seems like PCA by Anonymous Coward · · Score: 0

    I didn't read TFA but looks like they just discovered PCA http://en.wikipedia.org/wiki/Principal_component_analysis