Neural Network Drew Maps of 639,000 Galaxies express

Original: "Deep Learning Segmentation of Spiral Arms and Bars for 600,000 Galaxies in DESI"

· Ashley Spindler, Mike Walmsley, Karen L. Masters, Tobias Géron, Izzy L. Garland, B. D. Simmons, Jürgen J. Popp

arXiv:2606.16507 · 2026-06-15 · CC BY 4.0 · 1 min

The algorithm turned images of distant galaxies into detailed maps: marking spiral arms, central bars, and unexpectedly discovered rings.

An ordinary photo editor highlights objects in an image. The ZooBot:3D neural network works with distant galaxies like a cartographer drawing a map of an unexplored land. Thousands of volunteers manually traced spiral arms and bars (central stellar bands) on images, and from these lessons the algorithm learned to create not just lines, but heatmaps: each pixel tells whether an arm is there.

Surprise: the network started finding rings — rare circular structures. In the training examples, there were almost none. The algorithm seemed to figure out on its own what these 'donuts' look like.

With such maps, astronomers match data on stellar motion (spectroscopy) and brightness (photometry) to understand galaxy structure tens of times faster. The resulting catalog of 639,636 galaxies is open and already combined with the MaNGA and SAMI projects.

Whole graph →

Volunteers of the Galaxy Zoo project manually traced the outlines of thousands of galaxies, making millions of annotations. Now a single run of the neural network replaces this titanic effort.

Mapping galaxies is a key motif in science fiction, from Asimov’s 'Foundation' to 'Star Wars'; this work takes a step toward a real cosmic atlas.

Scientists

Fritz Zwicky Subrahmanyan Chandrasekhar Adam Riess John Archibald Wheeler Galileo Galilei William Herschel

Concepts

deep learning in medical imaging image classification desi tumor segmentation deep neural network spiral arms spiral galaxy medical imaging ai neural network galaxy mapping sdss

I am the author

Neural Network Drew Maps of 639,000 Galaxies express