Slashdot Mirror

← Back to Stories (view on slashdot.org)

New Algorithms Improve Image Search

Posted by ryuzaki0 on from the where's-waldo dept.

bc90021 writes "Electrical engineers from UC San Diego are making progress on an image search engine that analyzes the images themselves. At the core of this Supervised Multiclass Labeling system is a set of simple yet powerful algorithms developed at UCSD. Once you train the system (the 'supervised' part), you can set it loose on a database of unlabeled images. The system calculates the probability that various objects it has been trained to recognize are present, and labels the images accordingly. After labeling, images can be retrieved via keyword searches. Accuracy of the UCSD system has outpaced that of other content-based image labeling and retrieval systems in the literature. One of the co-authors works at Google, where the researchers have access to image collections at the largest of scales."

111 comments

  1. when I was your age by Anonymous Coward · · Score: 5, Funny

    I remember when we had to go to a gas station and *buy* porn. Now you have computers out there finding porn for you. You kids today have it too easy!

    1. Re:when I was your age by Prysorra · · Score: 2, Interesting

      That's one of the famous uses of image analysis - finding the presence of human skin in digital pictures.

      Skin detection.....5.5 million hits on Google.

      Once you can do this accurately, companies like McAffee and Norton can scan the internet and database pr0n sites for the whole web. Keep in mind that there's a subscription service that allows a Norton database to filter websites for them.

      Parents...

    2. Re:when I was your age by PPH · · Score: 3, Funny

      They chose the wrong name. It should have been "Supervised MUlticlass Tagging" or SMUT.

      --
      Have gnu, will travel.
    3. Re:when I was your age by risk+one · · Score: 1

      Uh, grampa... We'd really prefer if you didn't babysit anymore.

  2. Cool! by Deagol · · Score: 3, Interesting
    If this doesn't revolutionize the searching of online porn galleries, I don't know what will. :)

    Snarkiness aside, this is pretty cool stuff. I hope to see usable OSS code in a few years. Imagine how cool it would be to query "show me all pics with my daughter and her rabbits" and have it week through the 1000's of digital family photos.

    --
    Method of processing duck feet
    1. Re:Cool! by Cheapy · · Score: 5, Funny

      I find it disturbing that you combine porn, your daughter, and rabbits all in your post.

      You have issues.

      --
      Would you kindly mod me +1 insightful?
    2. Re:Cool! by UbuntuDupe · · Score: 2, Interesting

      Correct me if I'm wrong, and I'd like to be wrong, but isn't this (just) another application of Bayesian logic like is done for spam? They have some kind of way of quantifying the image in a number of variables and then use training to match certain variable values to a search term.

      (Even if it is, I don't want to trivialize the road from theory to practice, I just want to know what's different.)

      I did something a little while ago where I had a program search through text, and for all occurrences of all n-character strings (where you choose n) appearing, it would gather the information about how often each other character comes after each string appearing in the text. Then you'd give it an n-character string and it would use those probabilities to generate a new body of text. It was cool, even if it generated complete garbage except for large n.

      I hope to see usable OSS code in a few years.

      You mean for this algorithm, or at all?

      --
      Apology to Ubuntu forum.
    3. Re:Cool! by Tackhead · · Score: 4, Funny
      > If this doesn't revolutionize the searching of online porn galleries, I don't know what will. :)
      >
      > Snarkiness aside, this is pretty cool stuff. I hope to see usable OSS code in a few years. Imagine how cool it would be to query "show me all pics with my daughter and her rabbits" and have it week through the 1000's of digital family photos.

      ...the coolness of which is directly proportional to hotness of your daughter, the hotness of whom must then be further weighted by multiplying her hotness by some function of her age. The age-multiplier curve features an abrupt discontinuity that jumps 0.00 to 1.00 at age 18, and some sort of exponential backoff function that starts decreasing the multiplier at around age 35-45.

      But apart from the fact that it's almost Easter, what's with the rabbits? *clickity clic*-hey, I didn't know you could do that with Cadbury easter creme eggs!

      (Rule #34: There is porn of it. No exceptions.)

    4. Re:Cool! by andphi · · Score: 2, Funny

      I was going to assume that his daughter is little and likes rabbits because they're cute and fuzzy, or that she's somewhat older and keeps rabbits because they're cute, fuzzy, and more manageable than other animals. But, sadly this is Slashdot, so images that contain girls but aren't pr0n are apparently incomprehensible.

    5. Re:Cool! by chris_eineke · · Score: 1

      Rule #1: You do not talk about ___.
      Rule #2: You do NOT talk about ___.

      --
      "All you have to do is be fragile and grateful. So stay the underdog." Chuck Palahniuk, Choke
    6. Re:Cool! by Anonymous Coward · · Score: 5, Funny
      > But, sadly this is Slashdot, so images that contain girls but aren't pr0n are apparently incomprehensible.

      Fortunately, this is Slashdot, so discussions of pr0n that don't feature square-waves, multipliers, and exponential backoff functions are apparently incomprehensible too!

      (What are these "girls" of which you speak? I only remember Millie Amp... she was imaginary, skinny as a wire, but when her insulation got stripped, she stopped resisting, got really hot, and started to moan "ohm, ohm, ohm"?)

    7. Re:Cool! by Anonymous Coward · · Score: 0

      Hey, this sounds like a great opening to argue about Bayesian logic vs. fuzzy logic.

    8. Re:Cool! by Maxo-Texas · · Score: 1

      So the search algorithm is going to swear the girl was 18?

      --
      She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
    9. Re:Cool! by Anonymous Coward · · Score: 0

      Rule #35: If there isn't porn of it, there will be porn of it.

    10. Re:Cool! by andphi · · Score: 4, Funny

      By 'girls', I mean the limiting reagent in human reproduction. As a class of compounds, 'girls' are extremely common but somewhat volatile, so creating bonds with them is sometimes difficult. They are attracted to other similarly elusive compounds. Examples of these attracting compounds include 'Time', 'emotional vulnerability', and 'financial stability'.

    11. Re:Cool! by hoggoth · · Score: 2, Interesting

      > ...the coolness of which is directly proportional to hotness of your daughter, the hotness of whom must then be further weighted by multiplying her hotness by some function of her age. The age-multiplier curve features an abrupt discontinuity that jumps 0.00 to 1.00 at age 18, and some sort of exponential backoff function that starts decreasing the multiplier at around age 35-45.

      Hotness = BeautyFactor * SexyFactor * AgeHotnesseAdjustment
      AgeHotnessAdjustment = cos(2*(Age-18)/3.14159)

      Gives you maximum hotness at 18, falling slowly in the 20's, dropping rapidly after that.
      Also, some hotness under 18 (lets be realistic!) , but not too far under 18

      --
      - For the complete works of Shakespeare: cat /dev/random (may take some time)
    12. Re:Cool! by smaddox · · Score: 1

      I thought it was just some fetish I'd never heard of before, until I finally realized they were totally separate comments.

    13. Re:Cool! by crawly · · Score: 2, Interesting

      Umm no, that isn't what you want at all, that would give you a pretty horrible periodic function.

      Try something like this
      if age<18: AgeHotnessAdjustment = 0
      else: AgeHotnessAdjustment = 1/exp((Age-18)/20)

      --
      GCS/S d-x s+(+): a C++++$ UL+$ P+ L++$ !E--- W++@ N++>$ !o !K-- w++$ !O !M !V PS++>$ PE !Y PGP+ t+ 5++ X++ R tv b
    14. Re:Cool! by Anonymous Coward · · Score: 0

      hey, his daughter could be someone like this too:

      A little girl walks into a pet shop and asks with the sweetest little lisp, "Do you have widdle wabbits?"

      The shopkeeper gets down on his knees, so that he's on her level, and asks, "Do you want a widdle white wabbit or a thoft and fuwwy bwack wabbit, or maybe one like that cute widdle bwown wabbit over there?"

      She puts her hands on her knees, leans forward, and says in a quiet voice, "I don't fink my pet python weally gives a thit."

    15. Re:Cool! by sdfad1 · · Score: 1

      Yeah, most of the stuff I recently got interested in follow the same
      'framework' - given a set of data (doesn't matter what), you extract
      some features you are interested in and classify them. The features
      can have discrete values (sensor-A triggered, item B detected, test C
      positive), or continuous (humidity D = 90%, length E = 3.4 m,
      etc). Pass those feature vectors through a blackbox classifier, and
      attempt to 'fit' the features into a suitable class.

      The classes can be discrete (if binary, this could be a detection
      system: -1 means object not detected, 1 mean object detected, or test
      passed/positive etc, multivalued would be as explained in this
      example) in which case we have classification. If the classes were
      continuous, we have regression (given these features (age, weight,
      country), the most likely estimate for a person's height is 1.75
      m). When you design/train a classifier, you are basically looking for
      a function f(features) -> class that best describes your observed
      training data.

      So you see, it's all the same to me :). What is hard about images is
      in how smart you are in extracting those features. All you get is a
      bunch of pixels after all. The values they take on varies as lighting
      level, distance and colour varies. Additionally, objects in images can
      be rotated, reflected/mirrored, scaled, sheared, or distorted by the
      camera lens & position. Because of this, it's really difficult to
      classify pixels. It gets worse if all you are getting (as in this
      application) is low resolution highly compressed images (because of
      the additional artifacts).

      The Markov text generator you mentioned could be generalised to
      techniques dealing with N-grams. If you see for example Shannon's
      landmark paper about entropy - he talked about the very same technique
      for generating plausible looking paragraphs of text, so the Mark
      V. Shaney (ref wikipedia) algorithm has probably been around in some
      form or other since 'antiquity'.

      --- RANT ---

      As for Bayesian logic, yeah it's all cool, but there really is a HUGE
      class of techniques and algorithms one needs to draw upon to do AI. It
      takes forever to learn them all, but you almost need to know them all
      to be able to confidently claim that any technique is optimal for a
      particular situation... add to this the sometime fraudulent claims,
      myths and hype that has been circulating since the AI winter days, and
      you really have be careful. I'm a stauch Bayesian, and for the life of
      me could not get fuzzy logic. Why? There's the Cox axioms for
      reasoning with uncertainty, that forms the foundation for probability,
      something like probabilities must be greater than 0, must sum for
      mutually exclusive things, and I think there's probably something for
      products independent events - look it up. Thing is, there's a
      thing called the Dutch book that guarantees a loss to anyone who
      violates the Cox axioms in their reasoning. IE if you are playing the
      financial markets and you take two probabilities/beliefs/certainty
      factors/membership functions and use their min/max for and/or instead
      of products and sums (caveats apply), then doesn't this violate the
      Cox axiom? Doesn't that mean someone else using Bayesian inference
      would eat your lunch? Those fuzzy logicists always sidestep this by
      saying a fuzzy membership function is not a belief or
      probability. Well, it looks like one, it quacks like one, and it
      doesn't satisfy Cox's axioms. What gives?

      --- END OFFTOPIC RANT/CHALLENGE ---
      PS I thinks they're called Cox's axioms -> I am not a mathematician.

    16. Re:Cool! by Anonymous Coward · · Score: 0

      Rules 1 through 50:
      4chan sucks.

  3. so how does this ... by sarathmenon · · Score: 4, Funny

    change the way I search for Natalie Portman p0rn?

    --
    Microsoft: "You've got questions. We've got dancing paperclips."
    1. Re:so how does this ... by Anonymous Coward · · Score: 0

      The result would probably be less pictures of cows with flat utters.

    2. Re:so how does this ... by Anonymous Coward · · Score: 0

      It wouldn't, that little bitch isn't in any porn.

    3. Re:so how does this ... by physicsboy500 · · Score: 0

      you just draw a picture of her boobs and bam!!!
      NATALIE PORTMANIA!

      --
      The original generic sig.
    4. Re:so how does this ... by Rakshasa+Taisab · · Score: 1

      Yeah, and once we've tested it on that crap, let's make it search for _good_ porn...

      Damn... I thought geeks had good taste in porn. How ever did I manage to keep that illusion for so long while on /..

      --
      - These characters were randomly selected.
    5. Re:so how does this ... by Anonymous Coward · · Score: 0

      All *I* ever needed were search terms "grits" "hot" and "Portman". What else do you need, really? All else is superfluous.

    6. Re:so how does this ... by etherlad · · Score: 1

      Search terms: petrified, natalie portman, hot grits. Mix to taste.

      --
      Soylens viridis homines es
  4. She-male detection by Anonymous Coward · · Score: 0

    How well does it work for Porn, hopefully it will be able to differentiate between a she-male and female.

    1. Re:She-male detection by Anonymous Coward · · Score: 0

      This Sir must indeed be the Acid Test of the semantic image algorithms!

  5. Probability by DoofusOfDeath · · Score: 3, Interesting

    The system calculates the probability that various objects it has been trained to recognize are present,

    The probability is either zero or one, because whether or not the feature being sought is present is a state of nature. It would be more helpful to call this number the confidence that the feature is present.

    1. Re:Probability by AnonymousCactus · · Score: 2, Informative

      The probability isn't zero or one because the system doesn't have perfect knowledge and the probability is with respect to what the system 'knows'. Probability here is estimated based on the limited representation of the algorithm, so it's saying that based on the things I've seen before with similar features that were labeled 'tiger', X% were labeled 'tiger.' I would then expect this new thing to be a 'tiger' with a probability of X. (Exactly how they come up with their estimate is a bit more complicated :)) Confidence is a reasonable way of describing what that probability represents, but it's correct to say probability.

    2. Re:Probability by august+sun · · Score: 1, Interesting
      I though quantum mechanics allows for (mandates?) parallel universes for these variable states...

      So in Schroedinger's cat, in one universe the cat is alive and in one it is dead, and by observing the cat you only find out which universe you are in?

      Couldn't we therefore just say the probability is 1 that the object exists in some universe?

    3. Re:Probability by Anonymous Coward · · Score: 4, Insightful

      Not if it is a Bayesian probability.

    4. Re:Probability by aztec+rain+god · · Score: 1

      They should have clarified if they meant prior or posterior probability.

      --
      Sig cannot be found.
    5. Re:Probability by guycouch · · Score: 1

      You're not big on quantum superpositioning I take it.

    6. Re:Probability by zCyl · · Score: 1

      The probability is either zero or one, because whether or not the feature being sought is present is a state of nature.
      ... If you flip a coin, but don't look at the result, then the result is either heads or tails and the outcome present is a state of nature, but the PROBABILITY is 50% (for a fair coin). It is identical for images. It should tip you off when you find yourself using phrases like "the probability is either zero or one", because that statement identifies two states which each have a probability of happening in a random trial.
    7. Re:Probability by DoofusOfDeath · · Score: 1

      Good point. I concede.

    8. Re:Probability by $RANDOMLUSER · · Score: 4, Funny

      You're not big on quantum superpositioning I take it.
      I can take it or leave it.
      --
      No folly is more costly than the folly of intolerant idealism. - Winston Churchill
    9. Re:Probability by Anonymous Coward · · Score: 0

      "I can take it or leave it."

      Well I'm a big fan of quantum superpositioning, so I can take it AND leave it.

      Experiments on superpositioning having my cake and eating it too are forthcoming.

    10. Re:Probability by timeOday · · Score: 4, Interesting
      Or a fuzzy set, as (virtually) all set in the real world are.

      For instance, the set of pictures for which the statement "is this a picture of a chair" is true. There is no objective criteria for this. So imagine you have a bunch of pictures and show each one to a thousand people. Sometimes you might get 0 or 1000 "yes" responses, but often you'll get some number in between (because there are chairs, but barely visible, the picture includes a kids booster seat, or a rock big enough to sit on). This could be interpreted as a probability that somebody will consider a picture to be of a chair.

    11. Re:Probability by emlyncorrin · · Score: 4, Funny

      You're not big on quantum superpositioning I take it.
      I can take it or leave it. I can take it and leave it!
    12. Re:Probability by beav007 · · Score: 1

      Or a fuzzy set

      A handy system for searching images for cats.

      IM IN UR INTARNETS


      FINDN FUZZY PR0N!

    13. Re:Probability by Anonymous Coward · · Score: 0

      The system calculates the likelihood http://en.wikipedia.org/wiki/Likelihood_function that various objects are present...

    14. Re:Probability by CrimsonScythe · · Score: 1

      For instance, the set of pictures for which the statement "is this a picture of a chair" is true. There is no objective criteria for this. So imagine you have a bunch of pictures and show each one to a thousand people. Sometimes you might get 0 or 1000 "yes" responses, but often you'll get some number in between (because there are chairs, but barely visible, the picture includes a kids booster seat, or a rock big enough to sit on). This could be interpreted as a probability that somebody will consider a picture to be of a chair.

      I know there's a Ballmer joke in there somewhere. Did the depicted chair have a slight directional blur?

      --
      The view was horrible and the smell was even worse; Julie severely regretted becoming a proctologist.
    15. Re:Probability by Anonymous Coward · · Score: 0

      Or a conditional probability. E.g., of all images (in the sample of images known to the software) that contain each of three features x, y, and z, 40% are images of elephants, so there is a 40% probability that one such image is an image of an elephant. If the sample is representative of all images that will ever be presented to the software, then the 40% probability can be extended to images the system has never seen.

      Of course, if you train the system with wildlife images, and then turn it loose to classify astronomical images, it won't do very well.

  6. A military system I saw on a TV program ... by Anonymous Coward · · Score: 5, Interesting

    ... was similarly trained to recognise tanks in landscapes. I was doing really well - getting a great score on the fresh images it was presented with.

    Then they introduced it to a new batch of images and it fell apart.

    Turns out that the initial set of images had all the tanks shot on a sunny day and all the tankless images shot on a cloudy day (or vice versa). It had learned to tell a sunny day from a cloudy day.

    Ha ha.

    1. Re:A military system I saw on a TV program ... by Anonymous Coward · · Score: 0

      My teacher told the almost the same story. In his story, the system distinguished between allied and enemy tanks. Pictures of friendly tanks were taken on sunny days (because the photographers wanted the tanks to look good). Pictures of the enemy tanks, however, were shot with bad weather (because the photos were made by spies, they couldn't choose the weather conditions). So, yeah, the system could tell you if it was the time for picnic.

    2. Re:A military system I saw on a TV program ... by zolaar · · Score: 1

      No small feat, in and of itself?

      --
      One man's constant is another man's variable.
    3. Re:A military system I saw on a TV program ... by ClosedSource · · Score: 4, Insightful

      The system used neural nets. Generally you try NN's when you don't really understand the problem well enough to try a conventional approach. The problem with NN's is you really don't know what they are actually "learning".

    4. Re:A military system I saw on a TV program ... by Anonymous Coward · · Score: 0

      Yes, now you mention it I remember it was. Back then we thought NNs would fix everything. Soon learned not to trust everything they said on Tomorrow's World - CDs indestructible indeed!

      Heh, those crazy neurons!

    5. Re:A military system I saw on a TV program ... by Anonymous Coward · · Score: 0

      The system was a perceptron, which was the grandfather of artificial neural networks. It was the very first attempt to use a mathematical model of neurons, and we know today that on its own is pretty much useless.

    6. Re:A military system I saw on a TV program ... by Anonymous Coward · · Score: 0

      Yeah, I'm sure you "saw that on a TV program". It's funny that nobody can ever say what TV show they saw it on, or which article they read it in.

      However, I do agree that even in spite of this story almost certainly being apocryphal, it really could have happened. Not sure how relevant it is to this article, though, except that these guys also attempt to classify images. It's not exactly news that this doesn't always work flawlessly.

  7. Why is it better? by AnonymousCactus · · Score: 2, Informative

    I wish the article would mention more about why it is better than similar techniques that have been proposed in the past. (For example, http://luthuli.cs.uiuc.edu/~daf/papers/WAP-fin.pdf seems similar) For instance, where do they get their labels for the training data? A lot of people have tried using contextual words drawn from surrounding web text to limited success due to noise. It's also questionable how well their techniques can do if they need to pre-build a separate classification for each keyword. Finally, there are words that it seems impossible that they could ever distinguish. For example, 'man' vs. 'woman,' would be incredibly complicated for anything but a human. Where are the details? Oh yeah, it's a news story! Here's a link to the paper http://www.svcl.ucsd.edu/publications/journal/2007 /pami/pami07-semantics.pdf

    1. Re:Why is it better? by Tipa · · Score: 1

      Humans show a picture to the system and give it a list of elements present in the picture. That's where it gets the words.

    2. Re:Why is it better? by BurningPi · · Score: 0

      Finally, there are words that it seems impossible that they could ever distinguish. For example, 'man' vs. 'woman,' would be incredibly complicated for anything but a human. That makes it useless for pr0n searching; every time you searched for "naked woman", you would end up with goaste.
    3. Re:Why is it better? by nietpiet · · Score: 5, Informative

      I find it interesting which ones of the object-recognition and scene categorization algorithms make it to Slashdot.
      Why does this one make it?
      This is a very hot research topic at the moment.
      to name a couple of groups:

      http://www.robots.ox.ac.uk/~vgg/
      http://lear.inrialpes.fr/
      http://www.vision.caltech.edu/
      http://www.science.uva.nl/research/isla/
      http://www.cdvp.dcu.ie/
      http://www.informedia.cs.cmu.edu/
      http://www.research.ibm.com/slam/
      http://www.ee.columbia.edu/ln/dvmm/newResearch.htm

      oh, and people should not stare themselves blind on the claimed results.
      Research papers *always* have to present good results, or else you do not get published.
      Furthermore, these images are of a very high quality, make by professional photographers.
      Many algorithms perform very well on these ('corel'-like) sets, while utterly failing if applied on real-world data:
      http://www-nlpir.nist.gov/projects/trecvid/

    4. Re:Why is it better? by Coward+Anonymous · · Score: 1

      Actually, if I'm reading the paper correctly, the results aren't that spectacular. The hit rate appears to hover in the 30-35% range. It's apparently an improvement over other techniques but still nothing to form a company around. Hence the press release instead of another startup.

  8. Does it accually work... by BurningPi · · Score: 1, Funny

    ...or is it something like this?

  9. pr0n by __aapbzv4610 · · Score: 1

    Yes, personal searches were bound to be the first thing to be mentioned, but what about when others (ISPs, bosses, co-workers) are performing these searches on computers you use? I'm sure most people are smart enough not to do such things at work, but what about pop-ups (you couldn't help getting those kinds of popups while searching for a 'fix' to an app), false matches (boss doesn't view, only flags you if the keyword search comes back positive), etc?

    1. Re:pr0n by biscon · · Score: 1

      why would you boss care to search your computer for porn?

    2. Re:pr0n by LighterShadeOfBlack · · Score: 1

      Well it wouldn't be any different from existing text searches would it? Any kind of disciplinary action taken against an employee would need to be backed up with hard evidence. No company in the world would be dumb enough to try and take action without manually verifying it. And if they don't and you happen to be that employee just be glad: your ship has come in and it's manned by lawyers working on a no-win-no-fee basis.

      --
      Spelling mistakes, grammatical errors, and stupid comments are intentional.
    3. Re:pr0n by Anonymous Coward · · Score: 0

      Well it wouldn't be any different from existing text searches would it? Any kind of disciplinary action taken against an employee would need to be backed up with hard evidence. No company in the world would be dumb enough to try and take action without manually verifying it.

      Yeah, just like an RIAA/MPAA take down notice.

  10. Organizing my archive by Anonymous Coward · · Score: 0

    Cool! I can now have my image archive automatically sorted into my relevant categories:

    Blondes
    Brunettes
    Redheads
    Anal
    Animals
    Interracial
    Bukake

    But what if an image falls under more than category?

    1. Re:Organizing my archive by Feyr · · Score: 1

      tags

  11. Anyone tell Jeff Hawkins yet? by vrmlguy · · Score: 1

    I wonder if these students are using this software library?

    --
    Nothing for 6-digit uids?
  12. The problem is... by Life700MB · · Score: 4, Funny


    The problem is we all know what's gonna be the first result when searching "Caves on uranus"!!!

    --
    Great hosting 200GB Storage, 2_TB_ bandwidth, php, mysql, ssh, $7.95

    1. Re:The problem is... by Anonymous Coward · · Score: 0

      Google

      I'm feeling lucky

      No Problem

  13. The tech isn't mature enough yet by The+Orange+Mage · · Score: 4, Funny

    Run this story again when the system can tell the difference between D, DD, and DDD. Bonus points if it can handle "higher" criteria.

  14. autopr0n is dead by Gothmolly · · Score: 0

    long live autopr0n.

    Or will this be one of those absurd "hey there's skintones in this picture so we know its teh porno!" algorithms?

    --
    I want to delete my account but Slashdot doesn't allow it.
  15. Largest of scales? by Anonymous Coward · · Score: 0

    "One of the co-authors works at Google, where the researchers have access to image collections at the largest of scales."

    They just don't know the average slashdot user's porn collection yet.

  16. The first real success for a long sought after AI by Cafe+Alpha · · Score: 1

    An old AI joke was to call a limited domain version of this a "cat box". The idea a camera with a light on it that comes on whenever it's pointing at a cat.

  17. The Biggest Pornographer: +1, Interesting by Anonymous Coward · · Score: 0


    is The White House.

    I hope this help the R.I.C.O. case.

    Patriotically,
    K. Trout, C.P.A.

  18. Re:The first real success for a long sought after by DragonWriter · · Score: 1

    Is this really the first real success for that kind of "AI"? I'd rather thought that image classifiers based on neural networks and various other types of classifying techniques had been around for quite some time, and even used in realtime applications like self-driving cars that responded to road signs.

  19. Electrical engineers? by Anonymous Coward · · Score: 0

    Aren't electrical engineers supposed to be designing hardware? Why are they fooling around with image search and algorithms? Leave that to the computer science department and get back to work.

    1. Re:Electrical engineers? by djupedal · · Score: 1

      C'mon...get with the times.

      The keyword here is 'footprint', and since the new everything is now software, electrical engineers, as an example, would have a need to reduce the heat and energy 'footprint' of a given piece of equipment used to run the software that accomplishes the hit.

      If a hardware engineer can come up with a better search method via software that works quicker, reduces errors and further searching, uses a smaller processor, less heat, etc. then he/she has just reduced the energy footprint, regardless if he/she had to make the software that drove the smaller footprint or not.

    2. Re:Electrical engineers? by Anonymous Coward · · Score: 0

      As someone with an EE degree I can say without bias that EE's always write the best algorithms.

  20. Image SPAM by reh187 · · Score: 1

    It would be nice if we could humor these engineers to put as much effort into image spam filters :-P

    --
    Sarcasm is the recourse of a weak mind...
    --
  21. Not exactly new by denoir · · Score: 0, Flamebait
    Their work seems to be based on Gaussian Mixture Models which have been around for two decades or so. It's not a very advanced method either and there are a bunch of better adaptive systems for image recognition, wavelet neural networks being an obvious example (and they've been around more than a decade).

    There is absolutely nothing newsworthy about this. On the contrary, you'll find tons of similar works - mostly as senior year student projects in CS/AI.

    1. Re:Not exactly new by Anonymous Coward · · Score: 0

      Well, at least they're working on "an image search engine that analyzes the images themselves". This has got to beat previous approaches which concentrated mostly on the picture frame...

    2. Re:Not exactly new by denoir · · Score: 1

      Well, at least they're working on "an image search engine that analyzes the images themselves". This has got to beat previous approaches which concentrated mostly on the picture frame...
      Except for that people have already been doing that for decades. It is really sad to see that so many slashdotters are apparently 'AI' illiterate (if they are, imagine the general public).

      This 'news' is equivalent to reporting about researchers creating a new pointing device that looks like a small box with a cord and detects two-dimensional motion relative to its supporting service. The researchers are calling it 'the rodent'.

  22. Semantic Robot Vision Challenge at AAAI's by OtherFarm · · Score: 1

    a robot challenge that will test robots' vision and language understanding.
    the robots/sobots must be able to recognise objects automatically and perform tasks like: get the "star trek" poster or get the blue dry erase marker. the final event will be held at the twenty-second AAAI conference on artificial intelligence in vancouver, canada july 22-26 '07 [taken from ofpblog]

  23. Parent not just funny by EmbeddedJanitor · · Score: 4, Interesting

    Since a huge % (perhaps most) image searches are for porn, it is probably a worthwhile thing for a search server to quickly classify likely porn as a way to reduce search server loading.

    --
    Engineering is the art of compromise.
  24. Re:The first real success for a long sought after by Cafe+Alpha · · Score: 1

    Self driving cars? Yes, along with time-traveling Delorians and floating skateboards.

    Self driving cars have to be, (at least in recent years) an absolute con, just to get grant money. Would you trust technology as stupid as what we have?

  25. sounds like tagging , not image search by fikx · · Score: 1

    I keep waiting for a real image search to be created without the intermediarry step of tagging it with text.
    I'll be happy when I can tell the search page "find images like this" and give it an existing picture or a sketch. Tagging is too reliant on the consistant metadata to be useful in a general way. Humans can easily find all pictures of, say, fluffy the cat in a pile of photos from all different sources. Can we teach a computer how to do that without having to wait for it to re-tag images from different sources before it can search?
    Still, the better methods we find for tagging, the closer we get to that I guess...

    --
    AB HOC POSSUM VIDERE DOMUM TUUM
    1. Re:sounds like tagging , not image search by lukeinusa · · Score: 1

      Tagging images (as this system does) is in fact an effective way to do "real" image search. The work in Vasconcelos' lab and similar work on music annotation/retrieval UCSD's Computer Audition lab. has shown that representing an image as a distribution over semantic concepts (basically a "tag cloud") makes more reliable and accurate searches than using image- or audio-based feature comparisons alone. Basically, it's easier to find an image if you tell the computer "I'm looking for a white, fluffy cat" than if you say "I'm looking for a connected, amorphous set of white-ish pixels with a fluffy texture".

    2. Re:sounds like tagging , not image search by TuringTest · · Score: 1

      There is already an experimental search module that does what you describe - though it searched images in your hard drive only. I remeber seeing it advertised in /. a couple of years ago. Its accuracy leave something to be desired but it worked as proof of concept. The program was open source, I'm sure it'll still be available at freshmeat. Look for image galleries software, you'll find it there (I think it was either KMRML or imgSeek). Now it even has a web version.

      --
      Singularity: a belief in the "God" idea with the "demiurge" relation inverted.
    3. Re:sounds like tagging , not image search by zero_offset · · Score: 1

      KMRML! Of course! Now why couldn't I remember that name?

      --

      Slashdot quality declines as the number of hot grits posts decreases. - Provolt's Law, Apr-09-2005

  26. Re:The first real success for a long sought after by DragonWriter · · Score: 1

    Self driving cars? Yes, along with time-traveling Delorians and floating skateboards.


    There is a considerable difference between technology demonstrators and movie props.

    Neither tiltrotor transport aircraft nor warp drives are commercially available or in mass production, but they are in widely different categories. Self-driving cars are in the category with tiltrotors, while time-traveling Delorians (other than fixed-rate unidirectional travel, of course) are in the category with warp drives.

    Self driving cars have to be, (at least in recent years) an absolute con, just to get grant money. Would you trust technology as stupid as what we have?


    I don't trust humans as stupid as what we have on the road, for the most part.
  27. A combined solution might be useful by Sockatume · · Score: 1

    I recall seeing an image search a little while back where people tagged images manually, building up a weighted list of tags. It might be a good idea to use a system like that, to train a system like this. Like the spam filters we all know and love.

    --
    No kidding!!! What do you say at this point?
  28. Using games to get lots of tags by lukeinusa · · Score: 2, Insightful

    One complaint about this work is that it requires tagging an initial set of images that are needed to train the system. Vasconcelos' work uses the academic standard "Corel" dataset of labeled images but also uses tagged images from Flickr to train the system. Using human computation games like the ESP game for images and ListenGame www.listengame.orgfor audio, collecting data is not as tough as it once was...

  29. if... by RedElf · · Score: 1

    a picture's worth a thousand words, where are we going to store all the words for the useless myspace photo's that get archived?

    --
    You know, I have one simple request. And that is to have sharks with frickin' laser beams attached to their heads!
  30. How Many Non-techies Think This Isn't New? by wallywam1 · · Score: 2, Insightful
    It alwasy cracks me up on TV how they show a fingerprint or face recognition system searching for matches. The camera cuts to a computer that's looping away through a bunch of pictures until a match is found. Admittedly, I don't know all of the details, but obviously current systems must have some sort of indexed data points that are entered in a database and then you run queries against the database for potential matches.

    It's a little more plausible now that broadband is readily available but this has been portrayed on TV for years. Can you imagine some podunk field office connecting to an FBI database through a dialup and downloading high resolution images until they found just the right one? Then again, that would make for some good entertainment. Detective walks in..."I've got good news and bad news. The good news is we found the killer. The bad news is, he died of old age."

  31. I feel validated now... by pongo000 · · Score: 1

    ...because my master's thesis, back in 1996, was using neural nets with a fuzzy logic component to identify surface features on Landsat satellite imagery. The algorithm I came up with was even scale invariant.

    Guess I should have published and patented...damn...there goes any feelings of validation...

  32. Jim Gray by Dan100 · · Score: 0, Troll
    I wonder if this kind of technology could possibly locate Jim Gray's boat. As no wreckage was ever found, Gray was likely incapacitated in some way and the boat sailed under the autohelm far out into the deep ocean until the fuel tank ran dry. Technology like this could maybe scan satellite imagery looking for the Tenacious.

    I believe she's out there, somewhere.

  33. I might be missing the point, but ... by Anonymous Coward · · Score: 0

    they could just upload the "source" picture or data and wait for the remote machine to match it. Then download just the set of matching pics.

  34. Haar wavelets by neminem · · Score: 1

    Nobody's mentioned Haar wavelets yet? Weird.

    Look them up - they're part of OpenCV, and I'm pretty sure it's the same basic principles in action.

    1. Re:Haar wavelets by Anonymous Coward · · Score: 0

      It's a wavelet. Big whoop. Do wavelets classify images? Nope. Do nice, square, Haar wavelets do anything magical that all the others don't? Nope.

    2. Re:Haar wavelets by zero_offset · · Score: 1

      Actually denoir did mention wavelets a few posts back, and some moron modded it flamebait.

      --

      Slashdot quality declines as the number of hot grits posts decreases. - Provolt's Law, Apr-09-2005

  35. faggots by Anonymous Coward · · Score: 0

    you linux faggots will be looking for all that dick sucking porn. how about you fags just die off?

  36. Useful application by kn0tw0rk · · Score: 1

    This would be great for deviantArt, as one of the problems is mis-categorising their submissions. If a computer was able to help with that would make finding art of specific subjects/styles much easier.

    Please bring it on!

    --
    See my art -> http://herbevore.deviantart.com
  37. Re:The first real success for a long sought after by zero_offset · · Score: 1

    Self driving cars have to be, (at least in recent years) an absolute con

    VW is doing a pretty good job.

    http://www.dailymail.co.uk/pages/live/articles/new s/news.html?in_article_id=393401&in_page_id=1770

    --

    Slashdot quality declines as the number of hot grits posts decreases. - Provolt's Law, Apr-09-2005

  38. demonoid.com, mininova.org or btjunkie.org? by Ichimaru_Taishou · · Score: 1

    I missed the link, where can I torrent this?

  39. Finally! by Cheezymadman · · Score: 1

    Now I can search for porn stars that look like that girl in my Englsh class!

    --
    We're all going to die. i intend to deserve it.
  40. Cool! A Minnie Driver/Anne Hathaway love scene. by Impy+the+Impiuos+Imp · · Score: 1

    > The system calculates the probability that various objects it has been
    > trained to recognize are present, and labels the images accordingly

    "Ok, Joe. Let 'er rip on this new test database."

    Cock
    Cock
    Cock
    Vagina
    Cock
    Cock
    Hairy armpit

    "Oh, cool! The upgrade works and can distinguish it!"
    "Nah, wait until you see this!"

    Cock
    Cock
    Cock
    Midget with banana split in hairy ass crack with guy eating the banana split without using his hands on the Howard Stern show
    Cock
    Vagina
    etc.

    --
    (-1: Post disagrees with my already-settled worldview) is not a valid mod option.
  41. My image labeling tool by gr8dude · · Score: 1

    I once read about Google's image labeler, and decided to create a similar program, which would offer the same functionality, with additional features that are not available in Google's toy.

    The project does not have a name, it is described on my site - advanced image labeling tool. What makes it different is that besides collecting tags for an image, it also gathers other data about the tagger - age, sex, education, etc. My initial idea was to use it for various studies and establish connections between one's social status and the image labels they provide.

    Anyway, my point is that harvesting information about images can be fun, and it can have an impact on fields other than image processing or search engines.

    --
    The saddest poem
    1. Re:My image labeling tool by lukeinusa · · Score: 1

      There's a similar type of game at: http://www.likebetter.com/ This game uses your selection of images to infer facts about you, based on facts about others who like similar images. Pretty spooky...

  42. can it help me sort p0rn faster? by peter303 · · Score: 1

    So many boobs; so little time.

  43. Re:The first real success for a long sought after by Cafe+Alpha · · Score: 1

    Wow, it uses radar and "lazer" sensor instead of seeing. And it can evade road cones at high speed. That's the definition of safety isn't it?

    You should know what a useless toy that is.

    When they can trust their car to drive around schools, playgrounds, through ghettos, and New York city streets full of cars stopped in the middle and people behind them expecting them to break the law and go around (in traffic) call me.