SCARLET: Source separation in multi-band images by Constrained Matrix Factorization

Abstract

We present the source separation framework SCARLET for multi-band images, which is based on a generalization of the Non-negative Matrix Factorization to alternative and several simultaneous constraints. Our approach describes the observed scene as a mixture of components with compact spatial support and uniform spectra over their support. We present the algorithm to perform the matrix factorization and introduce constraints that are useful for optical images of stars and distinct stellar populations in galaxies, in particular symmetry and monotonicity with respect to the source peak position. We also derive the treatment of correlated noise and convolutions with band-dependent point spread functions, rendering our approach applicable to coadded images observed under variable seeing conditions. SCARLET thus yields a PSF-matched photometry measurement with an optimally chosen weight function given by the mean morphology in all available bands. We demonstrate the performance of SCARLET for deblending crowded extragalactic scenes and on an AGN jet -- host galaxy separation problem in deep 5-band imaging from the Hyper Suprime-Cam Stategic Survey Program. Using simulations with prominent crowding we show that SCARLET yields superior results to the HSC-SDSS deblender for the recovery of total fluxes, colors, and morphologies. Due to its non-parametric nature, a conceptual limitation of SCARLET is its sensitivity to undetected sources or multiple stellar population within detected sources, but an iterative strategy that adds components at the location of significant residuals appears promising. The code is implemented in Python with C++ extensions and is available at https://github.com/fred3m/scarlet

arXiv commentaccepted by Astronomy & Computing

References in corpus (27)

• Detailed Decomposition of Galaxy Images. II. Beyond Axisymmetric Models
• K-corrections and filter transformations in the ultraviolet, optical, and near infrared
• The Hyper Suprime-Cam SSP Survey: Overview and Survey Design
• Wide-Field InfrarRed Survey Telescope-Astrophysics Focused Telescope Assets WFIRST-AFTA 2015 Report
• The Dark Energy Survey: more than dark energy - an overview
• First Data Release of the Hyper Suprime-Cam Subaru Strategic Program
• The Hyper Suprime-Cam Software Pipeline
• GEMS: Galaxy Evolution from Morphologies and SEDs
• Shapelets: I. A Method for Image Analysis
• Galaxy Zoo : 'Hanny's Voorwerp', a quasar light echo?
• Structure of Disk Dominated Galaxies I. Bulge/Disk Parameters, Simulations, and Secular Evolution
• GALAPAGOS: From Pixels to Parameters
• The Galaxy Zoo survey for giant AGN-ionized clouds: past and present black-hole accretion events
• Galaxy Colour, Morphology, and Environment in the Sloan Digital Sky Survey
• Precision simulation of ground-based lensing data using observations from space
• The Symmetry, Color, and Morphology of Galaxies
• Dark Energy Survey Year 1 Results: The Impact of Galaxy Neighbours on Weak Lensing Cosmology with im3shape
• The Ellipticity Distribution of Ambiguously Blended Objects
• Sparse and Non-Negative BSS for Noisy Data
• Crowded Field Galaxy Photometry: Precision Colors in the CLASH Clusters
• Imaging Extended Emission-Line Regions of Obscured AGN with the Subaru Hyper Suprime-Cam Survey
• Multi-band morpho-Spectral Component Analysis Deblending Tool (MuSCADeT): Deblending colourful objects
• Nearly Optimal Sparse Fourier Transform
• NMF with Sparse Regularizations in Transformed Domains
• Nonnegative Matrix Factorization (NMF) with Heteroscedastic Uncertainties and Missing data
• Crowded Cluster Cores: Algorithms for Deblending in Dark Energy Survey Images
• Block-Simultaneous Direction Method of Multipliers: A proximal primal-dual splitting algorithm for nonconvex problems with multiple constraints

Cited by in corpus (55)

• Second Data Release of the Hyper Suprime-Cam Subaru Strategic Program
• Third Data Release of the Hyper Suprime-Cam Subaru Strategic Program
• The Dark Energy Survey Data Release 2
• Deblending and Classifying Astronomical Sources with Mask R-CNN Deep Learning
• Dark Energy Survey Year 3 Results: Measuring the Survey Transfer Function with Balrog
• Strong lens systems search in the Dark Energy Survey using Convolutional Neural Networks
• Photometry of high-redshift blended galaxies using deep learning
• Deblending galaxy superpositions with branched generative adversarial networks
• Deep Generative Models for Galaxy Image Simulations
• Cosmology with the Wide-Field Infrared Survey Telescope -- Synergies with the Rubin Observatory Legacy Survey of Space and Time
• Deblending galaxies with Variational Autoencoders: a joint multi-band, multi-instrument approach
• Cosmology dependency of halo masses and concentrations in hydrodynamic simulations
• Strong lensing in UNIONS: Toward a pipeline from discovery to modeling
• Optimising and comparing source extraction tools using objective segmentation quality criteria
• Imaging Extended Emission-Line Regions of Obscured AGN with the Subaru Hyper Suprime-Cam Survey
• Detection, Instance Segmentation, and Classification for Astronomical Surveys with Deep Learning (DeepDISC): Detectron2 Implementation and Demonstration with Hyper Suprime-Cam Data
• Weak lensing mass reconstruction using sparsity and a Gaussian random field
• Beyond Ultra-diffuse Galaxies. I. Mass--Size Outliers among the Satellites of Milky Way Analogs
• 21st Century Statistical and Computational Challenges in Astrophysics
• A model-independent characterisation of strong gravitational lensing by observables
• Bayesian photometric redshifts of blended sources
• The Vera C. Rubin Observatory Data Preview 1
• Deblending Galaxies with Generative Adversarial Networks
• Metadetection Weak Lensing for the Vera C. Rubin Observatory
• Gaussian Process Classification for Galaxy Blend Identification in LSST
• Galaxy blending effects in deep imaging cosmic shear probes of cosmology
• Generalised model-independent characterisation of strong gravitational lenses VIII. automated multi-band feature detection to constrain local lens properties
• PICZL: Image-based Photometric Redshifts for AGN
• Galaxy Deblending using Residual Dense Neural networks
• Beyond Ultra-diffuse Galaxies. II. Environmental Quenching of Mass-Size Outliers among the Satellites of Milky Way Analogs
• Deep Transfer Learning for Blended Source Identification in Galaxy Survey Data
• A Synthetic Roman Space Telescope High-Latitude Time-Domain Survey: Supernovae in the Deep Field
• Hybrid Physical-Deep Learning Model for Astronomical Inverse Problems
• Astrometric Jitter as a Detection Diagnostic for Recoiling and Slingshot Supermassive Black Hole Candidates
• Score-matching neural networks for improved multi-band source separation
• Gaussian Mixture Models for Blended Photometric Redshifts
• Disentangling transients and their host galaxies with Scarlet2: A framework to forward model multi-epoch imaging
• The End of Galaxy Surveys
• Ground-based image deconvolution with Swin Transformer UNet
• Joint survey processing: combined resampling and convolution for galaxy modelling and deblending
• Processing GOTO data with the Rubin Observatory LSST Science Pipelines I : Production of coadded frames
• Approximate Inference for Constructing Astronomical Catalogs from Images
• Deblending Overlapping Galaxies in DECaLS Using Transformer-Based Algorithm: A Method Combining Multiple Bands and Data Types
• DECam Survey for Low-Mass Stars and Substellar Objects in the UCL and LCC Subgroups of the Sco-Cen OB Association (SCOCENSUS)
• Neural Posterior Estimation for Cataloging Astronomical Images with Spatially Varying Backgrounds and Point Spread Functions
• Blending effects on shear measurement synergy between Euclid-like and LSST-like surveys
• TDE 2025abcr: A Tidal Disruption Event in the Outskirts of a Massive Galaxy
• A Bayesian Convolutional Neural Network for Robust Galaxy Ellipticity Regression
• Emulating redshift mixing due to blending in weak gravitational lensing
• Decomposition of stellar populations in CosmoDC2 galaxies using SCARLET and Deep Learning
• The Exploration of Local VolumE Satellites (ELVES) Survey: A Nearly Volume-Limited Sample of Nearby Dwarf Satellite Systems
• Effect of dust in circumgalactic haloes on the cosmic shear power spectrum
• Skykatana: a scalable framework to construct sky masks for the Vera Rubin Observatory and large astronomical surveys
• Virgo Filaments VI: H$\alpha$ clumps in the filaments around the Virgo galaxy cluster
• SpectralUnmix: A Torch-Based Regularized Non-negative Matrix Factorization