GOOGLE

How Google's RAISR tech helps photos load faster and save data

shutterstock_170404607

We all have experienced the frustration of having an erratic internet connection, when the time for a video to buffer or a picture to load takes forever. Not to mention the amount of data used up. Google is set to change that with its new machine-learning (ML) technology, Rapid and Accurate Image Super-Resolution (RAISR). It allows users to view high-resolution photos using 75 per cent lesser data bandwidth. Currently, RAISR works on Google+ only.

Let's assume that you are viewing photographs on someone's Google+ account. On a patchy internet connection, a high-resolution photograph would take more time to load and consume more data as well. With RAISR in action, you will see the same high-quality image loading faster, using lesser data than otherwise.

To do this, RAISR first fetches a low-resolution version of the photograph and uses its ML to convert it into a high-quality version, with a smaller file size.

John Nack, product manager at Google+, explains in a blog post how it works,"In practice, at run-time RAISR selects the most relevant filter from the list of learned filters to each pixel neighbourhood in the low-resolution image.” These filters are then applied to the low-quality image to recreate an image that would look exactly like the high-resolution version of the original image.

This also eliminates the need for users to upload low-quality photos.

RAISR Photo credit: https://www.blog.google/

Usually, when you reduce the file size of a high-resolution image, it results in a pixilated, blurry version that takes away the beauty of the picture. RAISR, having trained in machine learning and artificial intelligence, can work with low-quality versions (up to one-fourth of the original).


The developers at Google claim that RAISR trumps other methods of converting images to super-resolution. It is “roughly 10 to 100 times faster” and can run even on a mobile device in real-time.

The technology has been active for Google+ users since November 2016. This week Google announced through the blog post that the technology will roll out “more broadly” in the coming weeks.

“We're already applying RAISR to more than 1 billion images per week,” says Nack, “reducing these users' total bandwidth by about a third.”

He adds that because of RAISR, users have been able to display large images on Google+ while using up to 75 per cent less bandwidth for every image that it has been applied to.

This browser settings will not support to add bookmarks programmatically. Please press Ctrl+D or change settings to bookmark this page.
Topics : #Google | #technology

Related Reading