Cleaning Images with Gaussian Process Regression

Zhang, Hengyue; Brandt, Timothy D.

doi:10.3847/1538-3881/ac1348

Cited by 7 publications

(5 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In astronomy, GPR has already found numerous applications, including foreground removal (Mertens et al 2018;Ghosh et al 2020;Kern & Liu 2021;Soares et al 2021), foreground modeling by point-source interpolation (Pinter et al 2018), interpolating galactic-scale fields (Platen et al 2011;Yu et al 2017;Leike & Enßlin 2019;Williams et al 2022), cleaning/ interpolating missing data (Czekala et al 2015;Baghi et al 2016;Ndiritu et al 2021;Zhang & Brandt 2021), predicting photometric redshifts (Way et al 2009;Almosallam et al 2016), light-curve analysis (Evans et al 2015;Littlefair et al 2017;McAllister et al 2017;Angus et al 2018), radial velocity analysis (Haywood et al 2014;Barclay et al 2015;Rajpaul et al 2015), removing instrumental systematics (Gibson et al 2012;Junklewitz et al 2016), interpolating atmospheric turbulence to improve astrometry (Fortino et al 2021;Léget et al 2021), modeling PSF variations (Gentile et al 2013), or predicting object properties (Bu et al 2020;Fielder et al 2021).…”

Section: Introductionmentioning

confidence: 99%

“…The work of Fortino et al (2021) notably improves astrometry (measurements of the position of stars) on images from the Dark Energy Camera (DECam), which is the same camera used for validating our photometric model in Section 5. The extraction of clean foreground models by interpolating to remove point sources or bad pixels shown in Pinter et al (2018) and Zhang & Brandt (2021) serve as a proof of concept for this work. We further derive the correction to and uncertainties on the fluxes of those point sources as a result of the background model.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Photometry on Structured Backgrounds: Local Pixel-wise Infilling by Regression

Saydjari

Finkbeiner

2022

ApJ

View full text Add to dashboard Cite

Photometric pipelines struggle to estimate both the flux and flux uncertainty for stars in the presence of structured backgrounds such as filaments or clouds. However, it is exactly stars in these complex regions that are critical to understanding star formation and the structure of the interstellar medium. We develop a method, similar to Gaussian process regression, which we term local pixel-wise infilling (LPI). Using a local covariance estimate, we predict the background behind each star and the uncertainty of that prediction in order to improve estimates of flux and flux uncertainty. We show the validity of our model on synthetic data and real dust fields. We further demonstrate that the method is stable even in the crowded field limit. While we focus on optical-IR photometry, this method is not restricted to those wavelengths. We apply this technique to the 34 billion detections in the second data release of the Dark Energy Camera Plane Survey. In addition to removing many >3σ outliers and improving uncertainty estimates by a factor of ∼2–3 on nebulous fields, we also show that our method is well behaved on uncrowded fields. The entirely post-processing nature of our implementation of LPI photometry allows it to easily improve the flux and flux uncertainty estimates of past as well as future surveys.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Photometry on Structured Backgrounds: Local Pixel-wise Infilling by Regression

Saydjari

Finkbeiner

2022

ApJ

View full text Add to dashboard Cite

show abstract

“…Meanwhile, machine learning has also been applied to the field of nuclear physics [45][46][47][48][49][50][51][52][53][54][55][56][57][58][59][60]. As one of the machine learning algorithms, the Gaussian process has provided new ideas for the studies of many important physical problems in recent years [61][62][63][64][65][66]. The Gaussian process is a popular machine learning algorithm because it can provide error bars for the predictive values.…”

Section: Introductionmentioning

confidence: 99%

Theoretical predictions on α-decay properties of some unknown neutron-deficient actinide nuclei using machine learning *

Yuan¹,

Bai²,

Ren³

et al. 2022

Chinese Phys. C

View full text Add to dashboard Cite

Neutron-deficient actinide nuclei provide a valuable window to probe heavy nuclear systems with large proton-neutron ratios. In recent years, several new neutron-deficient Uranium and Neptunium isotopes have been observed using α-decay spectroscopy [Z. Y. Zhang et al., Phys. Rev. Lett. 122, 192503 (2019); L. Ma et al., Phys. Rev. Lett. 125, 032502 (2020); Z. Y. Zhang et al., Phys. Rev. Lett. 126, 152502 (2021)]. In spite of these achievements, some neutron-deficient key nuclei in this mass region are still unknown in experiments. Machine learning algorithms have been applied successfully in different branches of modern physics. It is interesting to explore their applicability in α-decay studies. In this work, we propose a new model to predict the α-decay energies and half-lives within the framework based on a machine learning algorithm called the Gaussian process. We first calculate the α-decay properties of the new actinide nucleus . The theoretical results show good agreement with the latest experimental data, which demonstrates the reliability of our model. We further use the model to predict the α-decay properties of some unknown neutron-deficient actinide isotopes and compare the results with traditional models. The results may be useful for future synthesis and identification of these unknown isotopes.

show abstract

“…All that can be done is to identify the cosmic ray by the spatial structure of the counts on the array (van Dokkum 2001;Pych 2004). The data must then be discarded or, if necessary, interpolated over (Zhang & Bloom 2020;Zhang & Brandt 2021).…”

Section: Introductionmentioning

confidence: 99%

Likelihood-based Jump Detection and Cosmic Ray Rejection for Detectors Read Out Up-the-ramp

Brandt

2024

PASP

View full text Add to dashboard Cite

This paper implements likelihood-based jump detection for detectors read out up-the-ramp, using the entire set of reads to compute likelihoods. The approach compares the χ 2 value of a fit with and without a jump for every possible jump location. I show that this approach can be substantially more sensitive than one that only uses the difference between sequential groups of reads, especially for long ramps and for jumps that occur in the middle of a group of reads. It can also be implemented for a computational cost that is linear in the number of resultants. I provide and describe a pure Python implementation that can process a 10-resultant ramp on a 4096 × 4096 detector in ≈20 s, including iterative cosmic ray detection and removal, on a single core of a 2020 Macbook Air. This Python implementation, together with tests and a tutorial notebook, are available at https://github.com/t-brandt/fitramp. I also provide tests and demonstrations of the full ramp fitting and cosmic ray rejection approach on data from the JWST.

show abstract

Cleaning Images with Gaussian Process Regression

Cited by 7 publications

References 23 publications

Photometry on Structured Backgrounds: Local Pixel-wise Infilling by Regression

Photometry on Structured Backgrounds: Local Pixel-wise Infilling by Regression

Theoretical predictions on α-decay properties of some unknown neutron-deficient actinide nuclei using machine learning *

Likelihood-based Jump Detection and Cosmic Ray Rejection for Detectors Read Out Up-the-ramp

Contact Info

Product

Resources

About