Neural networks promise sharpest ever images

February 22, 2017
The frames here show an example of an original galaxy image (left), the same image deliberately degraded (second from left), the image after recovery with the neural net (second from right), and the image processed with deconvolution, the best existing technique (right). Credit: K. Schawinski / C. Zhang / ETH Zurich.

Telescopes, the workhorse instruments of astronomy, are limited by the size of the mirror or lens they use. Using 'neural nets', a form of artificial intelligence, a group of Swiss researchers now have a way to push past that limit, offering scientists the prospect of the sharpest ever images in optical astronomy. The new work appears in a paper in Monthly Notices of the Royal Astronomical Society.

The diameter of its lens or mirror, the so-called aperture, fundamentally limits any telescope. In simple terms, the bigger the mirror or lens, the more light it gathers, allowing astronomers to detect fainter objects, and to observe them more clearly. A statistical concept known as 'Nyquist sampling theorem' describes the resolution limit, and hence how much detail can be seen.

The Swiss study, led by Prof Kevin Schawinski of ETH Zurich, uses the latest in technology to challenge this limit. They teach a , a computational approach that simulates the neurons in a brain, what galaxies look like, and then ask it to automatically recover a blurred image and turn it into a sharp one. Just like a human, the neural net needs examples - in this case a blurred and a sharp image of the same galaxy - to learn the technique.

Their system uses two neural nets competing with each other, an emerging approach popular with the machine learning research community called a "generative adversarial network", or GAN. The whole teaching programme took just a few hours on a high performance computer.

The trained neural nets were able to recognise and reconstruct features that the telescope could not resolve - such as star-forming regions, bars and dust lanes in galaxies. The scientists checked it against the original high-resolution image to test its performance, finding it better able to recover features than anything used to date, including the 'deconvolution' approach used to improve the images made in the early years of the Hubble Space Telescope.

Schawinski sees this as a big step forward: "We can start by going back to sky surveys made with telescopes over many years, see more detail than ever before, and for example learn more about the structure of galaxies. There is no reason why we can't then apply this technique to the deepest images from Hubble, and the coming James Webb Space Telescope, to learn more about the earliest structures in the Universe."

Professor Ce Zhang, the collaborator from computer science, also sees great potential: "The massive amount of astronomical data is always fascinating to computer scientists. But, when techniques such as machine learning emerge, astrophysics also provides a great test bed for tackling a fundamental computational question - how do we integrate and take advantage of the knowledge that humans have accumulated over thousands of years, using a machine learning system? We hope our collaboration with Kevin can also shed light on this question."

The success of the project points to a more "data-driven" future for astrophysics in which information is learned automatically from data, instead of manually crafted physics models. ETH Zurich is hosting this work on the space.ml cross-disciplinary astrophysics/computer-science initiative, where the code is available to the general public.

Explore further: Astronomers find faintest early galaxies yet, probe how the early universe lit up

More information: Kevin Schawinski et al, Generative Adversarial Networks recover features in astrophysical images of galaxies beyond the deconvolution limit, Monthly Notices of the Royal Astronomical Society: Letters (2017). DOI: 10.1093/mnrasl/slx008

Related Stories

A universe of 2 trillion galaxies

January 16, 2017

An international team of astronomers, led by Christopher Conselice, Professor of Astrophysics at the University of Nottingham, have found that the universe contains at least 2 trillion galaxies, ten times more than previously ...

Astronomers teach a machine to analyse galaxy images

July 8, 2015

A team of astronomers and computer scientists at the University of Hertfordshire have taught a machine to 'see' astronomical images. The technique, which uses a form of artificial intelligence called unsupervised machine ...

'Pandora's Cluster' seen by Spitzer

September 29, 2016

This image of galaxy cluster Abell 2744, also called Pandora's Cluster, was taken by the Spitzer Space Telescope. The gravity of this galaxy cluster is strong enough that it acts as a lens to magnify images of more distant ...

Recommended for you

Dawn of a galactic collision

December 14, 2017

A riot of colour and light dances through this peculiarly shaped galaxy, NGC 5256. Its smoke-like plumes are flung out in all directions and the bright core illuminates the chaotic regions of gas and dust swirling through ...

1 comment

Adjust slider to filter visible comments by rank

Display comments: newest first

Eikka
5 / 5 (1) Feb 23, 2017
The problem with neural networks is the same as having a guy with a bunch of photoshop filters clean the image: they'll see what isn't there and amplify it till it looks like a feature is present.

That's because the neural network doesn't care what's in the data, but what it looks like based on previous examples, so it's "inventing" details rather than bringing them out. That has horrible consequences for science if people take these images as real rather than just "artistic expressions" of a sort.

The real test is when it encounters a galaxy that doens't look like what it's been trained with. It's bound to clobber it up entirely, trying to shoehorn the "programmed imagination" and the actual picture together.

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

Click here to reset your password.
Sign in to get notified via email when new comments are made.