Researchers Develop Method for Getting High-Quality Photos from Crappy Lenses


There are many reason high-quality lenses cost as much as they do (and in some cases that is quite a lot), and one of them is that high-end lenses use many specially-designed elements that are perfectly-positioned to counteract aberrations and distortions.

But what if you could correct for all of that in post? Automatically? With just the click of a button? You could theoretically use a crappy lens and generate high-end results. Well, that’s what researchers at the University of British Columbia are working on, and so far their results are very promising.

The technique was presented at SIGGRAPH 2013, and it may some day provide a software alternative for those who can’t afford high-end glass. For their experiments, they developed a hand-made lens using only one element and then processed the resulting test images through their software to generate sharper results.

Check out their SIGGRAPH video below:

We won’t get into the technical bits (you can read the full paper here) but the basic premise is that once this software knows the point spread functions (PSFs) of your cheap lens, it can correct for blur, distortion and aberration and “recover” a high-quality image.

Here are some photos that show how their test images looked before (top) and after (bottom) sharpening with their computational imaging techniques:









The results are impressive, but for now there are still many hurdles left to jump before something like this could be brought to market. They have to figure out a way to account for the different PSFs of objects at different distances, and if the aperture gets any more open than f/2, the system runs into issues.

Still, this is a very promising start. To read more about the technique, check out the full paper. And if you’d like to see how their technique fared when using a non-homebrew lens — specifically, a Canon EF 28-105mm Macro — head over to this link for high-res samples.

(via Reddit)

Image credits: Photographs courtesy of the University of British Columbia

  • genotypewriter

    The manufacturers could store the calibration information in the lens itself after it is assembled and before it is packaged. That would solve the variations at assembly-time issue… but variations will still take place afterwards when things are bumped around. Maybe it’s cheaper to make better lenses than incorporate all these things… at least for cheap lenses.

  • genotypewriter

    But their $493289809483 equipment will also become better with this same method.

    Reminds me of FourThirds users (of the past, at leasT)… “but but… small sensor tech will catch up to APS-C and FF soon and there will be no real differences”… yeah, if APS-C and FF makers gave up on improving their sensors.

  • Zos Xavius

    If you ask me this isn’t very practical. Its far better just to have a good lens to begin with, though this does have the potential to remove all aberrations. But like you say, I wonder how long that stays calibrated. Lenses tend to get knocked around and elements tend to shift. Slight variations over the usage of the lens would make this all useless at the end of the day. Its cool technology. Don’t get me wrong. This could have a lot of potential in fixed lens cameras for sure.

  • jonquimbly

    Agreed on decisive moment- are they claiming it’ll work from a single photo? Oh yes, I see that’s implied but not clear in the article. Other considerations are bokeh, focused field quality, and ISO noise.

    I shouldn’t be so quick to dismiss it. When it comes along, I’ll definitely want to try it out.

  • genotypewriter

    The post-sale variations is inevitable and hence the reason why manufactures sort of imply that you need to send you lenses in every 2 years or so to be serviced, etc. At that time they can redo the calibration and store it back in the lens.

    Of course, at all times, the deconvolution should be an option rather than a forced thing… just like software vignetting compensation and CA compensation.

  • genotypewriter

    From what I gather, the multiple/local patterns are just used as different cases for approximating the function. The function should be the same for the entire image and will possibly have a variable ‘r’ to indicate how far along the radius a particular point/pixel is. When the function is properly optimised, it would generalise for all values of ‘r’ (i.e. every pixel basically), not just for the “known” cases.

    And forget zooms, the function would also need a ‘N’ variable for different f-numbers or else there will be too much “correction” applied for f-numbers between the two calibration f-numbers.

    With a zoom, things will suddenly become a lot more complicated. There will need to be an additional ‘f’ variable… and there will be parameter interactions… oh the joys of real-world function optimisation :)

  • genotypewriter

    Unsharp mask has been there since the 1930s in photo shops.

  • genotypewriter

    It’s not the same as stacking, obviously…

    “New reseach” doesn’t always mean “new outcome”… it could just mean “a new way”. If you’re in to research you’d know that each approach has its own pros and cons.

  • genotypewriter

    Like I said to someone else here, this is a different way of doing it… it’s still new research because of that. Every method has pros and cons and no one’s forcing anyone to use just one.

    For the record, you can’t “correct” CA, distortions, etc. At best you can hide their tell-tale signs like fringes and crooked geometry. The detail that was lost is gone forever. If you shoot the same scene side by side with a good lens and a regular lens that had its images “corrected”, you’d see the differences.

  • Alexander

    Great. Now how does that help me.

  • Olivier Lance

    Agreed. I think your comment is compatible with mine ^^

  • Sijimo

    Just did quick test with Sharpen filter with your test stills and my results look even better. I’ve been using it for years to restore details to slightly out of focus (“bad”) stills and videos.
    Nothing new to me. And Canon EF 50mm 1.8 is super cheap and good enough in most cases. I recall there was a development on sort of “shoot now, focus later technology”. That was interesting, this is not interesting.

  • Sijimo

    I take my words back about it being totally not interesting. I did a test again and it is slightly better then Photoshop’s Shake Reduction, or Smart Sharpen or Unsharp Mask if you watch it really close (at 400% zoom). But for 1080p video with H264 final delivery, it’s not that significant. Since after compression even at high bitrate most tiny details will be gone anyway.

  • jonquimbly


  • Richard Horsfield

    Law Enforcement CCTV dream software!

  • Igor Ken

    you are completely right, I meant something slightly different, but it doesn’t matter :)

  • TheNightstalker

    You guys didn’t get the point of what I was trying to point out. Those guys (you know who) will always come in, and say something along the lines of “well it will never look as good as my 50mm f/.95 Noctilux”. Sure, maybe it won’t, but that’s not the point of it. The point of it is to give the average consumer the ability to get great photos, without spending $10,000 on a single lens.

  • Igor Ken

    and it’s a great thing! :)