AI method to upscale low-resolution images to high-resolution

From small to not so pixel-perfect large
EnhanceNet-PAT is capable of upsampling a low-resolution image (left) to a high definition version (middle). The result is indistinguishable from the original image (right). Credit: Max Planck Institute for Intelligent Systems

Scientists at the Max Planck Institute for Intelligent Systems in Tübingen have used artificial intelligence to create a high-definition version of a low resolution image. While not pixel-perfect, the system produces a better result.

Technology to create a large-sized image from a low-resolution image is known as single-image super-resolution (SISR) technology. SISR has been studied for decades, but with limited results. Software adds extra pixels and averages them with the surrounding pixels, but the result is blurriness. Researchers at the Max Planck Institute of Intelligent Systems propose a new approach to give images a realistic texture when magnified from small to large using machine learning. Artificial intelligence is applied, and an adaptive algorithm for upsampling the image learns from experience to improve the result.

The learning process is much like that of a human. "The algorithm is given the task of upsampling millions of to a high-resolution version, and is then shown the original. Notice the difference? OK, then learn from your mistake," says Mehdi M.S. Sajjadi, who together with Dr. Michael Hirsch and Prof. Dr. Bernhard Schölkopf,, developed the EnhanceNet-PAT technology. Once EnhanceNet-PAT is trained, it no longer needs the original photos.

According to the researchers, the technology is more efficient than any other SISR technology currently on the market. In contrast to existing algorithms, EnhanceNet-PATdoes not attempt pixel-perfect reconstruction, but rather aims for faithful texture synthesis. By detecting and generating patterns in a low-resolution image and applying these patterns in the upsampling process, EnhanceNet-PAT adds extra pixels to the low-resolution image accordingly. For most viewers, the result is very much like the original photo.

Explore further

Automatic image retouching on your phone

More information: EnhanceNet: Single Image Super-Resolution Through Automated Texture Synthesis. arXiv.
Citation: AI method to upscale low-resolution images to high-resolution (2017, October 27) retrieved 18 September 2019 from
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.

Feedback to editors

User comments

Oct 27, 2017
The result is indistinguishable from the original image...

...except that it isn't. By a long way.

Oct 28, 2017
Anyone who can't see a BIG difference between the original and the AI version, should see an eye doctor immediately.

Oct 28, 2017
Well, perverts have gotten an early Christmas!

Oct 28, 2017
There is still quite a bit of difference but it's a still a lot better than the completely fuzzy one

Oct 29, 2017
If the center image is actually reconstructed from the image on the left, this is probably one of the best systems for enhancing images around. It still is a smeared version of the original though; but closer than one would expect.

Oct 30, 2017
What happens if you give it a picture outside of the training set?

I.e. what if it doesn't know that the picture contains this particular bird in this particular scene, but merely something that looks like it might be it. Would it try to reconstruct it the same way, and mess up the result?

Seems to me that given a truly random image, it would fail the same way the google deepdream fails by hallucinating dogs and cats and eyes everywhere.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more