I believe that everyone will encounter such troubles: when you get a good-looking photo, you will want to click in to see the big picture, but loading the big picture is very time-consuming, and it often fails to load.

However, a technology announced last November called Rapid and Image Super-(RAISR) could better address this problem.

(Image from:)

After learning, RAISR can generate high-resolution versions of low-resolution images. RAISR can output the same or better results than current image optimization techniques, and can be 10 to 100 times faster.

RAISR can save up to 75% of bandwidth while ensuring the resulting image quality is close to the original. It means that RAISR technology can process 1 billion pictures per week, which can save users about 1/3 of the bandwidth in total.

(Image from:)

The leftmost picture above is the original low-resolution picture, the middle is the result obtained by the traditional bicubic interpolation algorithm, and the right is the picture output by the RAISR algorithm. It can be seen that the results of the RAISR algorithm are clearer and more effective in detail processing.

(Image from:)

(Image from:)

The treatment of the old man's eyes is undoubtedly satisfactory. But the horse sculpture shows the inadequacies of the RAISR algorithm. The leaves on the soft background suddenly sharpen, strangely.

(Image from:)

Despite its drawbacks, RAISR has a unique trick. Generally, the image will be distorted after compression, and the resolution will be reduced. As shown on the left, the lines have been blurred into one piece. The bottom texture cannot be repaired using traditional linear algorithms, but RAISR's special learning method gives it a strong ability to remove distortion. The image on the right is the result of the output, the lines between the lines have been cleaned up a lot.

(The title picture comes from:)

Yesterday, + product manager John Nack wrote in a blog post that a small number of device users have already experienced the new technology when viewing high-resolution photos on +.

What the user loads is actually a compressed photo, but running the RAISR algorithm on the device can restore the details of the image.

While RAISR is currently only on +, at the end of the blog, John said they will be expanding the rollout of the technology in the coming weeks.

What do you guys think!