Parameter-Decoupling-Informed Learning for Superresolution: Laboratory Validation and Galaxy Imaging
Abstract
We present a parameter-decoupling-informed learning model for estimating subwavelength separations of point sources, without requiring prior knowledge of source properties, control over the sources, or continued access to their light signals. The method builds on our parameter-decoupled quantum superresolution estimation theory, which circumvents the need to estimate nuisance parameters such as partial coherence, brightness imbalance, and relative phase. Our convolutional neural network further accommodates practical imperfections, including background noise, photon loss, and centroid and orientation misalignment. We validate the method using synthetic and experimentally generated images, achieving separation-estimation fidelity above 82% for separations as small as approximately 0.075σ, where σ is the Gaussian intensity-PSF standard deviation. This corresponds to an illustrative separation of ∼13.5 nm when σ=180 nm, suggesting potential relevance to length scales addressed by advanced active-source microscopy. Applied to archival Hubble images of NGC~4486B, the model yields nuclear separations of 9.3-13.2~pc, consistent with the previously reported value of 12~pc. Comparisons using the same training and testing images demonstrate substantially higher fidelity than conventional direct data-driven learning in the tested cases. The method's robustness to varying source properties and imaging imperfections supports retrospective astrophysical image analysis and motivates applications in live-cell microscopy and quantum metrology where source control or renewed optical measurements are impractical.
Related
- Parameter-Decoupled Superresolution with Airy-Disk-Informed Learning: Application to NGC 4486B
- Non-Gaussian Parameter Bias: Formalism and Validation
- Multi-Axis Selective Decoupling Framework for Free-Floating Space Manipulators
- Decoupling Limit of Quiver Theories and the Angular Spectra of Extreme C-metrics
- Cosmological Evolution of Primordial Black Holes: UV/IR Decoupling and the KM3NeT 220 PeV Neutrino Anomaly
BibTeX
@misc{sajia2026parameterdecouplinginformed,
title = {{Parameter-Decoupling-Informed Learning for Superresolution: Laboratory Validation and Galaxy Imaging}},
author = {Abdelali Sajia and Bilal Benzimoun and James Qu and Pawan Khatiwada and Guogan Zhao and Gang Zhang and Xiao-Feng Qian},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.35449},
eprint = {2610.35449},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.35449},
note = {CC BY-NC-SA 4.0},
}
RIS
TY - UNPB
TI - Parameter-Decoupling-Informed Learning for Superresolution: Laboratory Validation and Galaxy Imaging
AU - Abdelali Sajia
AU - Bilal Benzimoun
AU - James Qu
AU - Pawan Khatiwada
AU - Guogan Zhao
AU - Gang Zhang
AU - Xiao-Feng Qian
PY - 2026
DA - 2026/10/26
PB - arXiv
DO - 10.48550/arxiv.2610.35449
UR - https://arxiv.org/abs/2610.35449
AB - We present a parameter-decoupling-informed learning model for estimating subwavelength separations of point sources, without requiring prior knowledge of source properties, control over the sources, or continued access to their light signals. The method builds on our parameter-decoupled quantum superresolution estimation theory, which circumvents the need to estimate nuisance parameters such as partial coherence, brightness imbalance, and relative phase. Our convolutional neural network further accommodates practical imperfections, including background noise, photon loss, and centroid and orientation misalignment. We validate the method using synthetic and experimentally generated images, achieving separation-estimation fidelity above 82% for separations as small as approximately 0.075σ, where σ is the Gaussian intensity-PSF standard deviation. This corresponds to an illustrative separation of ∼13.5 nm when σ=180 nm, suggesting potential relevance to length scales addressed by advanced active-source microscopy. Applied to archival Hubble images of NGC~4486B, the model yields nuclear separations of 9.3-13.2~pc, consistent with the previously reported value of 12~pc. Comparisons using the same training and testing images demonstrate substantially higher fidelity than conventional direct data-driven learning in the tested cases. The method's robustness to varying source properties and imaging imperfections supports retrospective astrophysical image analysis and motivates applications in live-cell microscopy and quantum metrology where source control or renewed optical measurements are impractical.
ER -