Neural Posterior Estimation for Cataloging Astronomical Images with Spatially Varying Backgrounds and Point Spread Functions

2025-02-28

AI Summary

Astronomical imaging faces challenges in source detection and photometric uncertainty quantification due to spatially varying sky background and point spread function (PSF). To address this, we propose the first neural posterior estimation (NPE) framework explicitly designed for spatially varying covariates, enabling end-to-end probabilistic source catalog generation. Our method introduces a semi-synthetic image generation paradigm that supports stochastic spatial sampling of both background and PSF; it integrates convolutional neural networks with variational inference to construct a learnable conditional posterior estimator. Evaluated on real Sloan Digital Sky Survey (SDSS) data, our approach significantly improves source detection accuracy, galaxy/star classification performance, and achieves well-calibrated photometric uncertainties—demonstrating robustness and practical utility under complex observational conditions.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: State Estimation
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Neural posterior estimation (NPE), a type of amortized variational inference, is a computationally efficient means of constructing probabilistic catalogs of light sources from astronomical images. To date, NPE has not been used to perform inference in models with spatially varying covariates. However, ground-based astronomical images have spatially varying sky backgrounds and point spread functions (PSFs), and accounting for this variation is essential for constructing accurate catalogs of imaged light sources. In this work, we introduce a method of performing NPE with spatially varying backgrounds and PSFs. In this method, we generate synthetic catalogs and semi-synthetic images for these catalogs using randomly sampled PSF and background estimates from existing surveys. Using this data, we train a neural network, which takes an astronomical image and representations of its background and PSF as input, to output a probabilistic catalog. Our experiments with Sloan Digital Sky Survey data demonstrate the effectiveness of NPE in the presence of spatially varying backgrounds and PSFs for light source detection, star/galaxy separation, and flux measurement.

Problem

• Neural posterior estimation for cataloging astronomical images with varying backgrounds.
• Addressing spatially varying point spread functions in astronomical image analysis.
• Improving accuracy in light source detection and flux measurement.

Innovation

• Neural Posterior Estimation for astronomical images
• Handles spatially varying backgrounds and PSFs
• Uses synthetic data for neural network training