Deep Learning for space-variant deconvolution in galaxy surveys
arXiv:1911.00443 · doi:10.1051/0004-6361/201937039
Abstract
Deconvolution of large survey images with millions of galaxies requires to develop a new generation of methods which can take into account a space variant Point Spread Function (PSF) and have to be at the same time accurate and fast. We investigate in this paper how Deep Learning (DL) could be used to perform this task. We employ a U-Net Deep Neural Network (DNN) architecture to learn in a supervised setting parameters adapted for galaxy image processing and study two strategies for deconvolution. The first approach is a post-processing of a mere Tikhonov deconvolution with closed form solution and the second one is an iterative deconvolution framework based on the Alternating Direction Method of Multipliers (ADMM). Our numerical results based on GREAT3 simulations with realistic galaxy images and PSFs show that our two approaches outperforms standard techniques based on convex optimization, whether assessed in galaxy image reconstruction or shape recovery. The approach based on Tikhonov deconvolution leads to the most accurate results except for ellipticity errors at high signal to noise ratio where the ADMM approach performs slightly better, is also more computation-time efficient to process a large number of galaxies, and is therefore recommended in this scenario.
References in corpus (23)
- Deep Convolutional Neural Network for Inverse Problems in Imaging
- An Augmented Lagrangian Approach to the Constrained Optimization Formulation of Imaging Inverse Problems
- A Method for Weak Lensing Observations
- Learned Primal-dual Reconstruction
- Understanding Deep Convolutional Networks
- Solving ill-posed inverse problems using iterative deep neural networks
- Gravitational Lensing Analysis of the Kilo Degree Survey
- CNN-Based Projected Gradient Descent for Consistent Image Reconstruction
- Generalized SURE for Exponential Families: Applications to Regularization
- Plug-and-Play Priors for Bright Field Electron Tomography and Sparse Interpolation
- Dark Energy Survey Year 1 Results: Weak Lensing Shape Catalogues
- GREAT3 results I: systematic errors in shear estimation and the impact of real galaxy morphology
- The Third Gravitational Lensing Accuracy Testing (GREAT3) Challenge Handbook
- LOFAR Sparse Image Reconstruction
- A SURE Approach for Digital Signal/Image Deconvolution Problems
- Joint Planck and WMAP CMB Map Reconstruction
- High Resolution Weak Lensing Mass-Mapping Combining Shear and Flexion
- Biases in, and corrections to, KSB shear measurements
- Space variant deconvolution of galaxy survey images
- Euclid: Nonparametric point spread function field recovery through interpolation on a graph Laplacian
- Sparse point-source removal for full-sky CMB experiments: application to WMAP 9-year data
- Deep Unfolding of a Proximal Interior Point Method for Image Restoration
- Soft Autoencoder and Its Wavelet Adaptation Interpretation
Cited by in corpus (27)
- General framework for cosmological dark matter bounds using -body simulations
- SLITronomy: towards a fully wavelet-based strong lensing inversion technique
- Image deconvolution and PSF reconstruction with STARRED: a wavelet-based two-channel method optimized for light-curve extraction
- Deep Learning-based galaxy image deconvolution
- Observational Constraints on Hybrid Scale Factor in f(Q,T) Gravity with Anisotropic Space-Time
- STARRED: a two-channel deconvolution method with Starlet regularization
- Superresolution Reconstruction of Severely Undersampled Point-spread Functions Using Point-source Stacking and Deconvolution
- Search for hybrid morphology radio galaxies from FIRST survey at 1400 MHz
- Semi-Blind Source Separation with Learned Constraints
- Deep learning-based deconvolution for interferometric radio transient reconstruction
- Cosmological implications of non-minimally coupled f(Q) gravity
- Ground-based image deconvolution with Swin Transformer UNet
- Improvement of cosmological constraints with the cross correlation between line-of-sight optical galaxy and FRB dispersion measure
- A morphological segmentation approach to determining bar lengths
- Curvaton and quantum gravity effect on the tensor to scalar ratio of the chaotic inflation
- Learning Lens Blur Fields
- Deblurring galaxy images with Tikhonov regularization on magnitude domain
- Galaxy Image Deconvolution for Weak Gravitational Lensing with Unrolled Plug-and-Play ADMM
- Independent interferometric observable confirms black hole shadow in M87*
- Galaxy Image Restoration with Shape Constraint
- Very-Long Baseline Interferometry Imaging with Closure Invariants using Conditional Image Diffusion
- ShapeNet: Shape Constraint for Galaxy Image Deconvolution
- Probing AGN duty cycle and cluster-driven morphology in a giant episodic radio galaxy
- Circular Mystery: Exploring Diffuse Emission Surrounding a Radio Galaxy with uGMRT and VLA Multiwavelength Observations
- Blind multi-frame deconvolution for the correction of space-variant blur in images
- Constraints on Energy Conditions in Gravity
- Joint multiband deconvolution for Euclid and Vera C. Rubin images