Become a fan of Slashdot on Facebook


Forgot your password?
Graphics Technology

Disney Algorithm Builds High-Res 3D Models From Ordinary Photos 80

Zothecula writes "Disney Research has developed an algorithm which can generate 3D computer models from 2D images in great detail, sufficient, it says, to meet the needs of video game and film makers. The technology requires multiple images to capture the scene from a variety of vantage points. The 3D model is somewhat limited in that it is only coherent within the field of view encompassed by the original images. It does not appear to fill in data"
This discussion has been archived. No new comments can be posted.

Disney Algorithm Builds High-Res 3D Models From Ordinary Photos

Comments Filter:
  • Time Saver (Score:2, Insightful)

    by Anonymous Coward

    This is great for scenery, it is amazing how much effort goes into the background scenery that no one will really pay attention to, but if you get it wrong everyone pays attention suddenly.

  • You just copy and pasted right before the punctuation mark for the summary?
  • by Anonymous Coward on Monday July 22, 2013 @12:59PM (#44352285)

    The 3D model is somewhat limited in that it is only coherent within the field of view encompassed by the original images. It does not appear to fill in data

    Just have the CSI boys zoom and enhance. C'mon guys, they've been doing this for years.

  • The technology requires multiple images to capture the scene from a variety of vantage points.

    So simply the big difference here seems to be synthesis of properly spaced stereo cameras by using cameras positioned anywhere.

    But it sounds less impressive than some of the stereo movies released from older 2D movies where there wasn't any additional
    cameras, and someone simply assigned depths to parts of the image and put CGI to work on it.

    • by interiot ( 50685 )
      3D scanning is really important. Whenever we figure out how to do it faster/cheaper/easier, that's important. 3D scanning is useful for all kinds of future activities, from the maker movement (3D printer + 3D scanner = 3D copier), to gaming (eg. Kinect), to driving (eg. DARPA Grand Challenge), to mobile devices (eg. Google Glasses).
  • Cheating (Score:4, Informative)

    by the eric conspiracy ( 20178 ) on Monday July 22, 2013 @01:01PM (#44352311)

    The technology requires multiple images to capture the scene from a variety of vantage points.

    That's cheating.

  • See this page []; the Campanile movie is from SIGGRAPH 97. How is Disney's tech different?
  • Yes, its fairly easy to build a 3d model given enough input and the right algorithms. Look at all the 3d scanner software that uses kinect and multiple frames to construct a higher resolution model.

    • Easy? Not at all. IIRC, to be able to theoretically get the model, no... let me try again: to even determine where your cameras are and how they are oriented, you need to be able to define something like 11 points in 7 photos.

      At that, that just gets you to the point of having N equations, N unknowns. It doesn't give you the answer. Nor does it account for lens distortion. Throw in lens distortion, and you have that many more unknowns, therefore that many more points you'll need to define.

      Having thought abou

      • You can roughly calculate camera extrinsic and intrinsic (including lens distortion) parameters from 3 world coordinates. (Assuming they're all in the photos)
        One presumes with the amount of threshold and flexibility the system needs, that'll be plenty.

        It's not that bad, just needs a lot of tweaking to get things to err, give decent persistent output

        • That's fine, if it's already known. But if it is unknown, then you've got a more difficult piece to chew on.

  • ...where construction crews are responsible for building our universe every second. Somehow our hero accidentally falls into that realm, and sees some whitespace or voids. One of the crewmembers tells him, when nobody goes there, they don't really have to build it. That when one of their crew is negligent, and forgets to put stuff where they're supposed to be; this explains those incidents when you could have sworn you placed that hammer on that table a second ago...

  • Affine (Score:5, Informative)

    by tmarthal ( 998456 ) on Monday July 22, 2013 @01:25PM (#44352599) Homepage

    The name of the algorithm is called 'affine reconstruction' and is a fairly well studied algorithm in computer vision. It is great that Disney and co. are releasing software to semi-automate the data input and reconstruction.

    • My question is, where is the software? All I saw on the Disney Research page was a Youtube video, and a link to white papers and links to download the original images. Just thinking how much I would love to take some old home movies and generate some semi-3D scenes from them. Could also be useful in 3D film conversions (although I thought that this was a similar approach to what they were already doing).

      • Re:Affine (Score:4, Informative)

        by Anonymous Coward on Monday July 22, 2013 @02:57PM (#44353655)

        Possibly not for those particular use cases, but there certainly is already freely available software to do the "structure from motion" reconstruction trick; e.g., vSFM [] -- an easy(FSVO)-to-use frontend for a couple of tools from different research projects.

        • Oh cool, thanks, someone mod that up! I may play with this some when I get home. Now this leads to an additional quesion - both this, and the software from Disney, seems to focus mainly on buildings. I guess I can try myself, but I wonder how well this works with people? It would also be nice if I can feed in a video, but I guess I can always take my video, feed it into some video editing software and export as a jpeg or bmp securence or something.

      • by Threni ( 635302 )

        I think you're confusing Disney with some other company which does open source graphics software...

  • by bradgoodman ( 964302 ) on Monday July 22, 2013 @01:28PM (#44352641) Homepage
    It's called 123D Catch. They have an iPhone app and everything... []

    • by gl4ss ( 559668 )

      It's called 123D Catch. They have an iPhone app and everything... []

      123D catch creates actual 3d models. in fact, any 123d catch demo is more impressive than this.. this is mostly useful for post process 3d effect to movies, it's one direction.

  • by future assassin ( 639396 ) on Monday July 22, 2013 @01:30PM (#44352659) Homepage

    I have a program from the mid 90's that I got from a book about VRML [] which would turn say buildings in photos into 3d objects. I think it was only a demo so never really tired it out to see if it worked.

  • I wonder if all those frames that stored say, John Wayne, could be used to create a fairly good 3D likeness. If not now, maybe soon. Also, who would own the rights to those performances?
  • ...use 4D light fields...

  • Did anyone else notice flashing images on the screen while playing the video?

    I can't get it to stop exactly on the image but it looks like some sort of white and red striped thing near 7 seconds in to the video.


    • by Phibz ( 254992 )

      I hate responding to my own comment but I finally got it downloaded and stepped through the video.

      It's just a person in the shot wearing a white and red striped shirt.

  • Wasn't this the first time? Photosynth demo.. []
  • Overlap the photos you're taking by 60% & look at them through a Stereoscope... you get 3D.
  • One problem with cheap and homebrew CNC machines is that the cutting head loses track of its own position. The feedback from CNC machines comes to the computer from sensors on the position screws of the various axes. These can lose their calibration. I keep thinking that with this kind of technology, a computer can "visually" determine the precise position of the cutting head and also of the material being cut. It can update the computer constantly on the exact shape of the remaining material, and then comp
    • by Richy_T ( 111409 )

      Usually stepper motors are used. These *can* lose track of their position but it is typically not an issue. It would mean that something had jammed or bound.

  • From Wikipedia:
    Released by MetaCreations Corp. in 1999, this application allowed users to create 3D models based on one or more photographs taken from various angles.

    Great program. Never understood why it died. Assume it was corporate hijinks.

  • by l3v1 ( 787564 ) on Tuesday July 23, 2013 @01:21AM (#44358401)
    In another news, the Sun is shining. I mean seriously, light-field based 3D reconstruction has been around for many years. Hell, even one of my colleagues has built a rotating table-based camera setup to capture images and create a full 3D model. Just google light fields 3D reconstruction or structure from motion and smell the coffee.

    Yeah, great news.
  • So art imitating life, or life imitating art? Just saw the Futurama episode Forty percent leadbelly [] in which Bender takes a 2-D photo of a guitar, and gets it duplicated by a 3-D printer. Seems we are in the future already!

"No, no, I don't mind being called the smartest man in the world. I just wish it wasn't this one." -- Adrian Veidt/Ozymandias, WATCHMEN