Planet detection down to a few λ/D: an RSDI/TLOCI approach to PSF subtraction
Bibliographic record
Abstract
Most current high contrast imaging point spread function (PSF) subtraction algorithms use some form of a least-squares noise minimization to find exoplanets that are, before post-processing, often hidden below the instrumental speckle noise. In the current standard PSF subtraction algorithms, a set of reference images is derived from the target image sequence to subtract each target image, using Angular and/or Simultaneous Spectral Differential Imaging (ADI, SSDI, respectively). However, to avoid excessive exoplanet self-subtraction, ADI and SSDI (in the absence of a strong spectral feature) severely limit the available number of reference images at small separations. This limits the performance of the least-squares algorithm, resulting in lower sensitivity to exoplanets at small angular separations. Possible solutions are to use additional reference images by acquiring longer sequences, use SSDI if the exoplanet is expected to show strong spectral features, or use images acquired on other targets. The latter option, known as Reference Star Differential Imaging (RSDI), which relies on the use of reference images that are highly correlated to the target image, has been ineffective in previous ground-based high contrast imaging surveys. The now >200 target reference library from the Gemini Planet Imager Exoplanet Survey (GPIES) allows for a detailed RSDI analysis to possibly improve contrast performance near the focal plane mask, at ∼2-7λ/D separations. We present the results of work to optimize PSF subtraction with the GPIES reference library using a least-squares algorithm designed to minimize speckle noise and maximize planet throughput, thus maximizing the planet signal to noise ratio (SNR). Using December 2014 51 Eri GPI data in the inner 100 mas to 300 mas annulus, we find no apparent improvement in SNR when using RSDI and/or our optimization scheme. This result, while still being investigated, seems to show that current algorithms on ADI+SSDI data sets are optimized, and that limited gains can be achieved by using a PSF archive.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.005 | 0.006 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".