Efficient PSF Modeling with ShOpt.jl: A PSF Benchmarking Study with JWST NIRCam Imaging
Authors:
Edward Berman,
Jacqueline McCleary,
Anton M. Koekemoer,
Maximilien Franco,
Nicole E. Drakos,
Daizhong Liu,
James W. Nightingale,
Marko Shuntov,
Diana Scognamiglio,
Richard Massey,
Guillaume Mahler,
Henry Joy McCracken,
Brant E. Robertson,
Andreas L. Faisst,
Caitlin M. Casey,
Jeyhan S. Kartaltepe
Abstract:
With their high angular resolutions of 30--100 mas, large fields of view, and complex optical systems, imagers on next-generation optical/near-infrared space observatories, such as the Near-Infrared Camera (NIRCam) on the James Webb Space Telescope (JWST), present both new opportunities for science and also new challenges for empirical point spread function (PSF) characterization. In this context,…
▽ More
With their high angular resolutions of 30--100 mas, large fields of view, and complex optical systems, imagers on next-generation optical/near-infrared space observatories, such as the Near-Infrared Camera (NIRCam) on the James Webb Space Telescope (JWST), present both new opportunities for science and also new challenges for empirical point spread function (PSF) characterization. In this context, we introduce ShOpt, a new PSF fitting tool developed in Julia and designed to bridge the advanced features of PIFF (PSFs in the Full Field of View) with the computational efficiency of PSFEx (PSF Extractor). Along with ShOpt, we propose a suite of non-parametric statistics suitable for evaluating PSF fit quality in space-based imaging. Our study benchmarks ShOpt against the established PSF fitters PSFEx and PIFF using real and simulated COSMOS-Web Survey imaging. We assess their respective PSF model fidelity with our proposed diagnostic statistics and investigate their computational efficiencies, focusing on their processing speed relative to the complexity and size of the PSF models. We find that ShOpt can already achieve PSF model fidelity comparable to PSFEx and PIFF while maintaining competitive processing speeds, constructing PSF models for large NIRCam mosaics within minutes.
△ Less
Submitted 18 June, 2024; v1 submitted 21 January, 2024;
originally announced January 2024.
ShOpt.jl: A Julia Package for Empirical Point Spread Function Characterization of JWST NIRCam Data
Authors:
Edward Berman,
Jacqueline McCleary
Abstract:
As astronomical data grows in volume and complexity, the scalability of analysis software becomes increasingly important. At the same time, astrophysics analysis software relies heavily on open-source contributions, so languages and tools that prioritize both performance and readability are especially valuable. Julia, with its just-in-time compiler and high level syntax, offers a compelling altern…
▽ More
As astronomical data grows in volume and complexity, the scalability of analysis software becomes increasingly important. At the same time, astrophysics analysis software relies heavily on open-source contributions, so languages and tools that prioritize both performance and readability are especially valuable. Julia, with its just-in-time compiler and high level syntax, offers a compelling alternative to traditional languages like Python or C.
In this paper, we outline ShOpt.jl, a new software package for point spread function (PSF) characterization written in Julia. ShOpt.jl features a number of performance optimizations, such as multithreading, the use of preconditioners, and the implementation of the memory-limited Broyden-Fletcher-Goldfarb-Shanno algorithm, as well as the flexibility to choose between principal component analysis, an autoencoder, and analytic profiles for PSF characterization. As observatories like the James Webb Space Telescope bring astrophysics into a new era of wide-field, high-resolution imaging, the challenges of PSF modeling become more pronounced. Tools like ShOpt.jl provide the community with a scalable, efficient, and accurate solution to these challenges, while also demonstrating the potential of Julia as a language that meets the demands of modern astrophysical research.
△ Less
Submitted 31 January, 2024; v1 submitted 29 September, 2023;
originally announced October 2023.
