digg

Join DiggAbout

Discover the best of the web!
Learn more about Digg by taking the tour.

Stanford researchers developing 3D camera with 12,616 lenses

eurekalert.org — If adigital camera saw the world through thousands of tiny lenses, each a miniature camera unto itself", you ’d get a 2-D photo, but you’d also get something potentially more valuable: an electronic “depth map” containing the distance from the camera to every object in the picture, a kind of super 3-D.

Bury
show profanity settings
expand all
  • ajimmykid, on 03/20/2008, -4/+5i saw a demo of this. you could basically take a photo, and then change the focus of the photo on your computer
  • ObeseEurotrash, on 03/20/2008, -10/+0this is a troll post
  • dwatten, on 03/20/2008, -5/+0this is gonna change the game, cant wait
  • Chordinator, on 03/20/2008, -4/+1Sounds made up to me.
  • lazersailer, on 03/20/2008, -1/+113...2...1... Youporn.com explodes.
    • cdigioia, on 03/20/2008, -0/+1Good point, my reaction to this news just went from "oh thats neat"
      to
      "OH MY GAAWWDD THAT'S AMAZING"
  • mopro, on 03/20/2008, -2/+7i'd rather just spend the money on 12,616 naked Tibetan hookers... i mean cookies. yea, cookies.
    • JudgeMonkey, on 03/20/2008, -0/+1yeah, I kinda scrolled through the article, and the word naked sorta popped out at me. Then the word probe... I was less interested.
  • cubbiesx, on 03/20/2008, -1/+6Did they mention something about aperture science? Well then this whole thing is a lie!!!
  • makkaveli19, on 03/20/2008, -0/+4imagine the pron possibilities
  • smurfsahoy, on 03/20/2008, -1/+2Hmm, you know a human can do the same thing, with 2 lenses that change. Why are we not learning a lesson from that here? I am not terribly impressed.
    • ajimmykid, on 03/20/2008, -1/+2because that would require using the lenses constantly. with the array of cameras, you can take the picture and then adjust focus after the photo has been taken.
      • smurfsahoy, on 03/20/2008, -0/+2No, the two lenses would just have to scan through all focusing distances while taking the picture. Then you could adjust the focus after the picture is taken.

        In combination with heuristics, this could do exactly the same thing for about 1/600th of the price. It's definitely possible. I have evidence that it's possible whenever I look at anything and see depth... In fact, you can show a human a still photo, and they can give you pretty good estimates of how far away everything is. Most depth info can be calculated with just ONE lens. Add in quick focal scanning and binocular vision, and a robotic camera could do a good enough job that it would approach the threshold of what humans could appreciate.
        • EtherGnat, on 03/20/2008, -0/+1"No, the two lenses would just have to scan through all focusing distances while taking the picture. "

          That takes time, making it useless for anything with motion. This system would be able to do the same thing in an instant.
          • smurfsahoy, on 03/20/2008, -0/+1You are forgetting the heuristics, which both during and after cut down on ambiguity constantly and virtually instantly (certainly fast enough for normal 30 frames per second video, since humans do this all the time, and we see a bit faster than that).

            And even if you want to be SURE to get motion, you only need one lens for every major focal distance, which there are not 1200 of. Humans' focal muscles (for the eye's lens) are not that precise. Anything past 20 feet is pretty much all one plane of focus as far as we are concerned, and closer than that, there's maybe a dozen planes that are worth distinguishing before the difference becomes so minute that it's just a huge waste of time and money.
          • EtherGnat, on 03/20/2008, -0/+1You're saying a lens can capture an image at every focal depth in 1/30 of a second?
  • blackdude, on 03/20/2008, -1/+2no pics? buried
  • DCJoeDogaswell, on 03/20/2008, -0/+2I'm sorry, I HAVE to say it

    Pics or it didn't happen
  • FTLJohnson, on 03/20/2008, -0/+6Much more info - Illustrated: http://www.news.com/8301-13580_3-9874436-39.html?% ...

    My guess is that this tech is being developed for "Stanley" the car.
    • Dested, on 03/20/2008, -0/+1Wow that picture looks buildable. Someone should try this with the ut3 engine.
  • flamov, on 03/20/2008, -2/+1It's over 9000!
  • Stoyanov, on 03/20/2008, -0/+1A very interesting read. In my opinion more money should be invested in science, rather than in developing weapons.
    • JudgeMonkey, on 03/20/2008, -0/+1Yep, then we can design all this badass stuff, only to have some impoverished waif from some neighboring country brandishing a slingshot to take it all away.
      • EtherGnat, on 03/20/2008, -0/+1To be fair, we ought to be able to significantly reduce military spending and still protect ourselves. We're currently spending about the same amount on defense as the rest of the world combined, and expenses continue to spiral out of control. For example when adjusted for inflation the war in Iraq is likely to cost twice as much as World War I and 10 times as much as the first gulf war.
  • jeffchuck, on 03/20/2008, -0/+1The article fails to mention how this is an improvement over using two lenses to calculate depth. Humans are quite good at doing the job with two eyes, as the article mentions:

    "just as the left eye of a human sees things differently than the right eye"

    It also mentions that two lenses can capture 3D:

    "a camera with two lenses (or two cameras placed apart from each other) can take more interesting 3-D photos."

    So what's the benefit of over 9000 extra lenses?
  • robotto, on 03/20/2008, -1/+1Finally we can have a true 3D Starwars where we can look around and into Leia's cans
  • asjk, on 03/20/2008, -0/+1"Or the sensor could be deployed naked, with no objective lens at all. By placing the sensor very close to an object, each micro lens would take its own photo without the need for an objective lens. It has been suggested that a very small probe could be placed against the brain of a laboratory mouse, for example, to detect the location of neural activity." I'm not sure I follow how the last sentence is connected to the rest of this paragraph.
  • krenzo, on 03/20/2008, -0/+9This would be a big help for compositors working in the special effects industry. By having the depth recorded with the color information, you no longer need green screen backgrounds because you can instantly separate the foreground subject that you want to keep from the background that you want to replace. If you want to place a computer generated object in a scene, it will automatically be obscured by foreground objects yet remain in front of the background.
  • 223Sniper, on 03/20/2008, -0/+0yea thats nice. what were we talking about again?
  • xdevit, on 03/20/2008, -0/+1Porn will never be the same..
  • freeferty, on 03/20/2008, -0/+1Cool technology but only 1 pic on the page??
  • malchihi, on 03/20/2008, -0/+3Just remember our vision is currently just in 2D. Although many of you will argue that our vision is 3D, we use our 2D vision along with shadows and depth to perceive three dimensions. Just take a minute to think about it. If you look out in front of you the thing you see is a panoramic 2d image with a perceived third dimension.
    • xMedic, on 03/20/2008, -0/+1I always thought of it as: your brain, knowing the distance between your eyes and the angles of each eye when looking at something, can "calculate" distance to the object you're looking at.
  • wire02, on 03/20/2008, -0/+1Its good to see rule 34 alive and well in these comments
  • steakfry, on 03/20/2008, -0/+2I wonder how this compares to Ren Ng et al's microlens-based plenoptic camera that allows post-exposure digital refocusing (that they were working on in 2005, also at Stanford)...

    Abstract of paper:

    "This paper presents a camera that samples the 4D light field on its sensor in a single photographic exposure. This is achieved by inserting a microlens array between the sensor and main lens, creating a plenoptic camera. Each microlens measures not just the total amount of light deposited at that location, but how much light arrives along each ray. By re-sorting the measured rays of light to where they would have terminated in slightly different, synthetic cameras, we can compute sharp photographs focused at different depths. We show that a linear increase in the resolution of images under each microlens results in a linear increase in the sharpness of the refocused photographs. This property allows us to extend the depth of field of the camera without reducing the aperture, enabling shorter exposures and lower image noise. Especially in the macrophotography regime, we demonstrate that we can also compute synthetic photographs from a range of different viewpoints. These capabilities argue for a different strategy in designing photographic imaging systems.

    To the photographer, the plenoptic camera operates exactly like an ordinary hand-held camera. We have used our prototype to take hundreds of light field photographs, and we present examples of portraits, high-speed action and macro close-ups."

    Example pics & vid here:

    http://graphics.stanford.edu/papers/lfcamera/
  • Vektuz, on 03/20/2008, -0/+1or just do what google did and do the same thing but instead of millions of lenses, use a high speed scanning laser. They've already got a model that can do 640x480 at like 30fps, with 3d info for each pixel... and it doesnt cost a fortune.. ;p
    • steakfry, on 03/21/2008, -0/+1Does that actually measure luminance, or is it just a shape model? I'd love to check that out if you've got a link!
  •  

    Login or register to add a comment

    Create a new account or login to join in the conversation on Digg. You'll also be able to Digg stories to help promote things you like.


© Digg Inc. 2008 — Content posted by Digg users is dedicated to the public domain.
DIGG, DIGG IT, DUGG, DIGG THIS, Digg graphics, logos, designs, page headers, button icons, scripts, and other service names are the trademarks of Digg Inc.