DIY Photography

Hacking Photography - one Picture at a time

  • News
  • Inspiration
  • Reviews
  • Tutorials
  • DIY
  • Gear
Search

 
@diyphotography
Instagram

Submit A Story

New neural network repairs damaged and low-quality images

Share30
Tweet56
+1

December 4, 2017 by Dunja Djudjic Leave a Comment

We’ve seen some of the algorithms that can enhance low-quality photos. The researchers from Oxford University and the Skolkovo Institute of Science and Technology in Moscow have developed a new approach for restoring damaged or low-quality images. Instead of training the neural network with thousands of photos, their system called Deep Image Prior works everything out from the original image. And without any previous learning, it turns a pixelated or damaged photo into a hi-res one.

Dmitry Ulyanov, one of the co-authors of the research, explains that the “network kind of fills the corrupted regions with textures from nearby.” Instead of using the data from the datasets, Deep Image Prior redraws a blurry or damager picture until it gets it right. According to Interesting Engineering, some images turn out to be even better than the original input.

Despite the good overall results, Ulyanov admits that there are still failures in the redrawing attempts:

“The obvious failure case would be anything related to semantic inpainting, e.g. in-paint a region where you expect to be an eye — our method knows nothing about face semantics and will fill the corrupted region with some textures.”

The researchers mention some of the possible legal and ethical applications of this method. It can be used to filter out noise or upsample low-res photos, which photographers might find useful. Deep Image Prior can also be applied to the restoration of old and damaged photos, which museums and archives are likely to find useful.

But there’s a downside to this neural network as well. Other than restoring damaged, blurry, grainy and otherwise corrupted photos, Deep Image Prior can also remove the text placed over the images. This raises copyright concerns, considering that the algorithm could make image theft easier than ever.

Another question is whether apps like this could leave photo restorators without work. Sure, human brain and skill are still better than computers, so this certainly won’t happen today. But at some point in the future when the algorithms are perfected, the photo restorators’ jobs may become less wanted and less paid.

I think we can definitely view discoveries like this from both sides. There’s no doubt there can be very useful, but also harmful if they get into the wrong hands.

You can view more image examples here and read the full paper on this link.

[Deep Image Prior via Interesting Engineering, Inverse]

Share30
Tweet56
+1

Related posts:

New algorithm “enhances” low-res photos till they are tack sharp in hi-res Shutterstock’s tech lets you play with text on canvas to find photos. Not there yet Ostagram uses neural networks to create mind-bending images Facebook now lets you send 4K images in Messenger

Filed Under: news Tagged With: Enhance, hi-res, neural network, Neural Networks

« How to process and grade drone photos in Lightroom
5 small editing tricks that can make a big difference to your video »

Trending on DIYP

  • Wedding venue catches fire while filming the newlyweds’ first dance
  • The world’s first single shot 8×10 large format digital camera can be yours for a mere $106,000
  • This impossible long take was shot with a tiny drone in a Japanese high school
  • First look at the Pentax K1 Mark II and DFA 50mm f/1.4 lens
  • 2018 Pulitzer Prize winner for photography leaves his job and now works for a brewery

Recent Comments


Previous Polls

Dunja Djudjic is a writer and photographer from Novi Sad, Serbia. You can see her work on Flickr, Behance and her Facebook page.

John Aldred is based in Wishaw, Scotland and photographs animals in the studio and people in the wild.

You can find out more about John on his website or follow him on Facebook and YouTube.

JP Danko is a commercial photographer based in Toronto, Canada. JP
can change a lens mid-rappel, swap a memory card while treading water, or use a camel as a light stand.

To see more of his work please visit his studio website blurMEDIAphotography, or follow him on Twitter, 500px, Google Plus or YouTube.

JP’s photography is available for licensing at Stocksy United.

Clinton Lofthouse is a Photographer, Retoucher and Digital Artist based in the United Kingdom, who specialises in creative retouching and composites. Proud 80's baby, reader of graphic novels and movie geek!
Find my work on My website or follow me on Facebook or My page

Recent Posts

  • Capture One 11.1 update offers new workflow improvements and camera support
  • Canon launches Ivy; a rechargeable pocket-sized mini photo printer
  • Venus Optics soon to announce Laowa 10-18mm f/3.5-4.5 zoom lens for full frame cameras
  • Slum pictures – let’s talk about ethics
  • How to make your own DIY portable green screen with PVC pipe and a pillow case

Copyright © DIYPhotography 2006 - 2018 | About | Contact | Advertise | Write for DIYP | Full Disclosure