'

Artificial Intelligence can now be used to sharpen pixelated images and video

Artificial Intelligence can now be used to enhance pixelated images and video

Artificial Intelligence can now be used to sharpen pixelated images and video

Artificial Intelligence can now be used to enhance pixelated images and video

 

When ordinary people discuss artificial intelligence it usually turns into an "end or the world" discussion, and while intelligent machines could eventually lead to that the benefits/applications of artificial intelligence/deep learning are a lot wider than creating robotic servants or creating convincing chatbots. 

Magic Pony, a UK start-up that has been created by graduates of Imperial College London, has utilized machine learning techniques in order to process visual information, allowing them to process and improve images and videos in real time.  

The company has created a way to create high-resolution images from low-resolution ones, allowing them to improve the quality of video on the fly using ordinary graphics processors, allowing even mobile phones to enhance the image quality of videos and pictures. 

One way to use this new technology is to apply it to a streamed video, which is often heavily compressed in order to stream it to your PC over the internet and use GPUs to enhance the quality and clarity of the final image, especially when looking at text. 

Below is a quote from Rob Bishop, stating that this technology can currently be used with fast mobile GPUs, meaning that we may see some form of this technology applied to current video streaming platforms like YouTube and Twitch in the future, as it can work with modern mobile hardware. 

 

Online video-streaming businesses rely heavily on video compression,

Our first product demonstrates that image quality can be greatly enhanced using deep learning, and fast mobile GPUs now allow us to deploy it anywhere.

 

Artificial Intelligence can now be used to sharpen pixelated images and video   

 

The technology works because they have designed a computer that which when fed an image it converts it into a low-resolution version and  then compares the differences between it and the original image. This machine then uses the information collected over thousands of images to learn patterns that will allow it to enhance pixelated images.  

Above and below are images that have been enhanced using this technology, with the below image from a video stream showcasing a huge gain in visual quality, especially when looking at text and numbers onscreen. 

Right now the company states that they are in talks with several interested parties that wish to license this technology, though it is worth noting that this tech can be applied to a lot more than streaming online video. 

 

Artificial Intelligence can now be used to sharpen pixelated images and video

(Original Compressed Video (Left), Enhanced Video (Right)) 

 

One example that the company gave was that this technology could be used to dynamically create high-resolution textures for games and VR experiences using lower resolution examples or simply to generate miles of unique terrain textures using previously used examples. 

This technology could also be used to clean up the images taken from low-quality cameras or taken in low light environments, making this technology appealing to a large number of industries.  

 

You can join the discussion on Artifical intelligence being used to improve the quality to low resolution images and video on the OC3D Forums.  

 

«Prev 1 Next»

Most Recent Comments

16-04-2016, 15:10:13

thekeemo
Enhance!Quote

16-04-2016, 15:25:44

WYP
Quote:
Originally Posted by thekeemo View Post
Enhance!
CSI were telling the future all along?Quote

16-04-2016, 15:51:18

NeverBackDown
Now this is some good ol' innovation. Some great work and has a LOT of application uses.Quote

16-04-2016, 19:10:13

Kushiro
Just to enhance the guy on the side too

If only it could isolate which parts to enhance.Quote

16-04-2016, 20:33:55

Kleptobot
Would be interesting to see if you could analyse it to find which patterns it recognised the fastest and then create a video compression algorithm to compress those features the most and see how much compression you could achieve and still have good looking video ectQuote
Reply
x

Register for the OC3D Newsletter

Subscribing to the OC3D newsletter will keep you up-to-date on the latest technology reviews, competitions and goings-on at Overclock3D. We won't share your email address with ANYONE, and we will only email you with updates on site news, reviews, and competitions and you can unsubscribe easily at any time.

Simply enter your name and email address into the box below and be sure to click on the links in the confirmation emails that will arrive in your e-mail shortly after to complete the registration.

If you run into any problems, just drop us a message on the forums.