Become a fan of Slashdot on Facebook


Forgot your password?

CSI Style Zoom Sees Faces Reflected In Subjects' Eyes 103

mikejuk writes "A recent paper by Dr Rob Jenkins of the Department of Psychology at York University (UK) has managed to prove that you can get useful images of faces from the reflections in eyes. It really is as simple as zooming in. The catch is that the experiments were done with a 39 mega pixel camera — even so the actual final images were low resolution. In the experiment a number of people were photographed with a 'bystander' in a position so that a reflection of their face would be captured in the eye. The resulting extracted image of the reflection in the eye was only 27x36 and then rescaled using bicubic interpolation to 400x240 or bigger and enhanced using standard PhotoShop operations to normalize the contrast and brightness. Test subjects were able to match faces using the low resolution images but the important result was that if the subject knew the person in the photo then recognition went up to 90% with false positives down at 10%. So the next time you appear in a photo consider the fact that a simple procedure might reveal who you are with."
This discussion has been archived. No new comments can be posted.

CSI Style Zoom Sees Faces Reflected In Subjects' Eyes

Comments Filter:
  • Facebook tagging (Score:5, Insightful)

    by Overzeetop ( 214511 ) on Saturday December 28, 2013 @10:24AM (#45805039) Journal

    Coming to NSA-approved monitoring sites like Facebook: automatic detection of people you are with, even if they're not "in' the photograph!

  • Enhance! (Score:4, Funny)

    by Anonymous Coward on Saturday December 28, 2013 @10:24AM (#45805041)

    Gratuitous enhance []!

  • CP (Score:2, Troll)

    So we should be giving really high res cameras to pedophiles. That way the photos they take will have more than fear in the eyes of their victims.
  • by Anonymous Coward

    I picked up a Sony A7R in november and noticed that I could clearly see myself in the eyes of every person I photographed. At 36 mega pixels, consumer cameras like the A7R (and that nikon whos name I can't remember) are more than capable of the CSI effect.

    • by gl4ss ( 559668 )

      ..but it really isn't the csi effect.

      because csi effect refers to making up a face from let's say 3x3 grid of pixels - from something that it is clear to anyone(or should be) that there isn't enough data to make the zoom work.

      it just doesn't work! now making a shoddy picture from 30x30 grid or so, yeah, but false positives galore with that too.

      so for me, the study sounds just like they did a study about how to get headlines before doing the study, I bet a shitload of studies have been performed on low res

  • count of the cameras you are snapping pics with. I can only think of a couple of cameras that get 39 megapixels, and they are in the Pureview line of Nokia's products. Even those don't produce 39 megapixel images, they use a sampling method to make better images at ~9 megapixel final resolution.
    • by Anonymous Coward

      I'd recognize my hubby's ex even in a 1x1 image. :P

    • Problem is what is high res this year is mainstream next and in the cheapie aisle a year after that. When I started playing with cameras 0.03 MP was the norm and 2 MP cost an arm and a leg,heck even my 2011 phone is 3.2 MP . Now its getting really hard to find anything under 3 MP and 10 MP is becoming mainstream, next year I wouldn't be surprised to see 12 MP or higher become default.
      • by Znork ( 31774 )

        The pixel count increases as it's a selling point. The optics on most things like phones seem to stay the same utter crap, barely enough to saturate the ccd in full sunlight, leaving most of those megapixels as random noise. Leaving any CSI style zoom as something achievable in photos specifically taken with the proper equipment and setup to allow CSI style zoom but not in the junk quality images they pretend to enhance in CSI.

    • If you've only been exposed to cell phones and retail-level cameras you might think it's just a gimmick Nokia is using,'d be wrong. The Nikon D800 DSLR is at 36; Hasselblad/Mamiya/Pentax/Leica medium-format bodies are anywhere from 50 to 80. Have been since 2006 or so. Give it another two generations on the DSLR front and you'll be getting a $499 50MP Nikon or Canon at Wal-Mart.

      Anyway, even on my D5100@16MP I can pull recognizable images if it's a frame-filling portrait, anything more is just maki

      • And to get everything in focus at that level of detail you need a _very_ flat subject, or to have everything you want at the infinity focus level.

        The depth of field (area that is in acceptable focus) decreases as required detail goes up, not to mention that if you up the required detail enough you wind up being forced to use a larger aperture otherwise diffraction effects will be the limiting factor.. using a larger aperture decreases the depth of field even more.

        There are real practical limits for getting

    • by Rei ( 128717 )

      Couldn't one use super-resolution techniques [] from a video or series of images to increase the effective resolution?

    • by dwater ( 72834 )

      The Nokia 808 could produce images with almost 42MPixels :

      "Nokia 808 has a 41.3-megapixel CMOS FSI image sensor, 1/1.2-inch image sensor format with 7728 x 5368 pixels. Depending on the aspect ratio chosen by the user, it will use 7728 x 4354 pixels (33.6 Mpx) for 16:9 images, or 7152 x 5368 pixels (38.2 Mpx) for 4:3 images when using the default camera software. The commercial app CameraPro[12] can use the whole sensor size for capturing full 7728 x 5368 pixels (41.48 Megapixel) pictures with more than 30

  • by Anonymous Coward on Saturday December 28, 2013 @10:39AM (#45805129)

    As a photographer, I have to clean up the eyes and all the other reflective materials from my images, if any, or find myself in sweet spots where it doesn't happens.

    It's also common knowledge that you can actually learn about the photographer's lighting setup and materials by looking at the shadows AND light reflection in the eyes.

    There are also nice non-shopped photographies with that reflectiveness property being applied to produce interesting results, with inverted sceneries being shown in the eyes. Take the picture close enough and you won't have 50 pixels, but 500 of reflection. Take it with a high resolution camera and you'll get 1000-2000 pixels worth of goodness.

    So there's a direct correlation between a pixel count and the number of pixels for a detail.

    In other words, it's known that you get good reflectivity in eyes and it's known that using high resolution cameras you can get good details from eyes, inferring you can tell who's on the other side. Well ... yep. Next thing they'll figure out is by having a high sensitivity camera you can get cleaner pixels of the people themselves in the reflectivity, as the lighting is seldom as adequate and the contrast is much less prevalent in these reflections, so by having much cleaner 32 native bits per color, you could technically have a "perfect" rendition of the user, especially at high F

    Surprise me next time :) I'd even say that if you could have a consistent reflective model of a person's anatomy and a map of its clothings, you could potentially infer a person's surroundings, as every material is slightly reflective / refractive too. But we're really not there yet. :)

  • by wonkey_monkey ( 2592601 ) on Saturday December 28, 2013 @10:48AM (#45805183) Homepage

    A recent paper by Dr Rob Jenkins of the Department of Psychology at York University (UK) has managed to prove that you can get useful images of faces from the reflections in eyes.

    Didn't anyone think this couldn't be done?

    Are you sure he doesn't work in the Department of Bleedin' Obvious?

    • Didn't anyone think this couldn't be done?

      Duh. Did. Did anyone think this couldn't be done?

    • by AmiMoJo ( 196126 ) *

      The problem is that now some professor had demonstrated it every prosecution that lacks real evidence but "knows" the guy is scum and really wants to nail him will have another kind of dubious but impressive forensic evidence to use. Juries love this CSI shit, because science can't be wrong, right?

  • by flanktwo ( 1041494 ) on Saturday December 28, 2013 @10:58AM (#45805215)
    These guys at Columbia did something similar years ago, but used eyeball reflections and a cornea model to figure out what the person was looking at: []
    • And at least 30 years ago, I read of a guy who was studying classic paintings to see if he could discern anything in the reflections in painted eyes. Wold be even lower "res", but perhaps a careful artist did so without realizing it. Have no idea if anything came of it.

  • Not very practical (Score:4, Interesting)

    by vadim_t ( 324782 ) on Saturday December 28, 2013 @11:20AM (#45805327) Homepage

    Anybody with a recent DSLR can test that this can be done. I recently took a portrait if myself in my cat's eye.

    The trick to this though is that you need a DSLR with fairly high resolution, a good sharp lens, and have the photo be a closeup of the subject. None of which are features of the vast majority of security camera footage.

    If somebody was willing to spend amounts in the range of $1000 per camera, yes, this might be a possibility. Provided the person stared right into the camera while standing at a meter or so from it.

    Also, it'll probably stay this way. There are limitations to the useful resolution that can be achieved, so it's not possible to simply put a 1000 MP sensor into a security camera and suddenly be able to perform the tricks shown in CSI.

    • You are a bit off on the price. This one is $17,000 [].

    • by fatphil ( 181876 )
      And if you throw 10000 notes at the problem, you can take a photo where both the eye itself, which takes up a large proportion of the frame it is so close and the reflected scene about 1-2m away are in perfect focus, and with vivid colours to boot. One of the best photos I've seen in the last few years, and when my mate says "no photoshop apart from when gamut-mapping from raw to jpeg", I believe him. I'm still not sure I know exactly how he did it so brilliantly, I'm guessing it was very careful lighting.
    • by epine ( 68316 )

      So the next time you appear in a photo consider the fact that a simple procedure might reveal who you are with.

      Yeah, I hang out all the time in public spaces with KH-11 prosumer cameras concealed behind 1970s ceiling tiles in every room and corridor.

      It's this increasingly common tag line on the article submissions that makes Slashdot news for slack-jawed mooncalves.

    • by dwater ( 72834 )

      Not possible with a Nokia 808 (or at least using it's h/w) - for $350 USD? []

  • I am trying to figure out the newsworthy part in this - it's been known for ages that you can get details out of reflxions and high-def images...

    For the practical implications - we're not going to get any benefit from users' twitpics, blog images and fb photos, as those are rarely ever uploaded in full high-def+highest-quality, 41MP camera notwithstanding: I'd like to see

    • the resolution and quality actually used by the camera (user setting rarely push the camera to its max capcity to save on storage)
    • the re
  • How cool would this be if it were useful for anything at all?

    So the next time you appear in a photo consider the fact that a simple procedure might reveal who you are with."

    Uhh, whoever I was with? You mean whoever took the photo? You mean whoever posted the photo? wtf?

  • by MacTO ( 1161105 ) on Saturday December 28, 2013 @11:57AM (#45805531)

    ... the technology to convert unfocussed VGA resolution photographs and convert them into focused 39 megapixel photographs. That is why this technique is so useful on CSI.

  • The only novel thing, and by far the spookiest, about this "research" is that it's coming from a "Department of Psychology".

    As for the ramifications, it'll be a long time before North Korea's press officers get much overtime on the back of it. I doubt it'll be of interest even to domestic security services. However cheap and widespread hi-res cameras become, 90% reliability is woefully low. They can get much better than that already, just with a board and a bucket.

  • by Anonymous Coward

    I can get xrays off peoples eye reflections.

    I'm already so more future they can't even imagine.

  • Test subjects were able to match faces using the low resolution images but the important result was that if the subject knew the person in the photo then recognition went up to 90% with false positives down at 10%.

    So, if I know who I'm looking for I can find them in a blurry, low-res picture?

    How can anyone be expected to recognize anyone they don't know?

    Wouldn't the reflection in a subject's eye be of the picture taker with their camera in front of their face?

    • by lxs ( 131946 )

      How can anyone be expected to recognize anyone they don't know?

      Even without context that's a tricky one.

    • The implication is that for a human observer extracting a description would be hard but pattern matching is easy.

    • by dwater ( 72834 )

      perhaps they were also given regular photos of people - that's what I presumed by 'match faces'.

  • How long until the much-maligned faceback app [] in 'The Other Guys' becomes real?

  • []

    (based on Blade Runner of course, which did it first)

  • by damaki ( 997243 ) on Saturday December 28, 2013 @12:52PM (#45805871)
    Thank you for bringing up such quality content. You know, 39 megapixel pictures are only usable if you have high end glass on your camera, which only rich camera enthusiasts can afford. That excludes forever smartphone cameras (yeah, even those whatever megapixels Nokias), security cameras, small digital cameras, which would actually render useless 99% of all cameras on earth for that usage. So you only need to have one of these rich guys take a picture of a crime scene, yeah, only that...
    • by bytesex ( 112972 )

      It's a bit of shame that you feel the need to react so bitterly to an experiment whose real-life ramifications will be upon you soon enough. 39 megapixel is expensive today - will it be in five years?

      • by damaki ( 997243 )
        39 usable megapixels will always be expensive. High end lenses are damn expensive, and if your lenses do not have the required resolution, you're better off using 16 megapixels. For $1000, glass included, you won't get more than 16 real megapixels. And if you do not believe me, have a look at [] .
  • "...but the important result was that if the subject knew the person in the photo then recognition went up to 90%"

    So it's just like the photos of the guy with the stolen Debit card at the ATM.
    Just their mother would recognize them on those images.

  • So is it true a dead person's retina retains the image seen at the moment of death?

  • That was no ordinary camera they used for this...

    If you read the original paper ( []), buried in the detailsis the one where they used a Hasselblad H2D, which is a medium-format digital camera with an enormous 36.7mm x 49mm CCD.

    This camera is about 7 years old ( []), so it might be possible to find it for cheaper on the secondary market. The current version, the H5D, a

  • I need to start working on those goatse contact lenses!
  • Medium format film runs into 100s of "megapixels" (i.e. up to 10x the resolution of the digital camera they used) and large format even higher.
  • Now the question is: can we engineer the thing and create a fake? We can even imagine a post-processing filter in cameras to remove face from eyes in the picture. We already remove red eyes.
  • Everytime i came across a stories like this one, I remember a story here on Slashdot a long time ago, maybe in another time-space continuum, before 2001, maybe in 1999, about someone optical searcher in MIT who can create very HD images from staking multiple lowres image from a movie camera. This software was open source and I try it with y own images from small sony handycam with great result. I try to find reference to this software and research without any succes since. and I obvioulsy don't have the sf
  • ...the subject is looking at a picture..?

All seems condemned in the long run to approximate a state akin to Gaussian noise. -- James Martin