Neural Network Drew Maps of 639,000 Galaxies express
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.
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.
Original abstract
as written by the authors arXiv:2606.16507
We present a catalogue of segmentation maps identifying the extent of spiral arms and bars of 639,636 galaxies in the DESI Legacy Survey. To produce these maps, we have trained a deep U-net-style neural network using the pixel masks from the Galaxy Zoo: 3D citizen science project. The resulting data products are "soft" segmentation maps, which show the confidence of the model that a pixel lies within the spiral arms or bars of a galaxy. In this paper we detail the sample selection from DESI-LS using the machine classifications from Galaxy Zoo: DESI, the architecture of the U-net model--dubbed ZooBot:3D. We demonstrate the ability of the model to identify spiral arms and bars in a wide range of face-on disks, and identify an emergent ability to identify rings--despite only a small number of ring-type galaxies being present in the training data. Finally, we discuss the practical application of these data products to photometric imaging and IFU spectroscopy. The ZooBot:3D dataset is available for use publicly, and contains the full catalogue presented in this paper, along with cross-matched subsamples for the MaNGA and SAMI IFU surveys.
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.