Enhanced micrometre spatial resolution and usable photon flux for hybrid pixel detectors via deep-learning-based X-ray localization
Abstract
Micrometre spatial resolution is achieved for X-rays with the MÖNCH detector, a charge-integrating hybrid pixel detector with a 25 μm pixel pitch, by exploiting charge sharing between neighbouring pixels. The conventional analytical interpolation method, however, relies on a global mapping derived from the charge-sharing statistics and is applicable only to isolated single-photon clusters, which limits both the achievable resolution and the usable photon flux. To overcome these limitations, we trained deep learning models to localize single-photon and pile-up events based on high-fidelity simulation data. For knife-edge measurements at 12 keV and a low occupancy of 0.74% (photons per pixel and frame), where single-photon events dominate, the proposed method reaches a of the edge spread function of 1.36 μm, compared with 1.54 μm for the conventional eta interpolation. We demonstrate experimentally that resolving up to three photons per cluster raises the reconstruction efficiency from 59% to 93% at a 2.4{\times} higher occupancy of 1.81%, while preserving a of 1.39 μm.
Related
- Photon Calibration Techniques for High Resolution Cryogenic Detectors
- Comprehensive study of timing resolution in plastic scintillator detectors with wavelength-shifting fiber and silicon photomultiplier readout
- Improved aluminum film transport properties and enhanced TES response in athermal phonon detectors: implications for quasiparticle dynamics
- High resolution 3D localization of ultrasound sparse sources using a SAGE algorithm
- Tip-enhanced molecular fluorescence microscopy with atomic-scale resolution
BibTeX
@misc{xie2026enhanced,
title = {{Enhanced micrometre spatial resolution and usable photon flux for hybrid pixel detectors via deep-learning-based X-ray localization}},
author = {Xiangyu Xie and Anna Bergamaschi and Martin Brückner and Maria Carulla and Tomáš Celko and Roberto Dinapoli and Simon Ebner and Simone Emiliani and Khalil Daniel Ferjaoui and Erik Lars Fröjdh and Viveka Gautam and Dominic Greiffenberg and Shqipe Hasanaj and Viktoria Hinger and Marius Hürst and Vadym Kedych and Thomas King and Shuqi Li and Carlos Lopez-Cuenca and Leonid Lunin and Federica Marone and Alice Francesca Mazzoleni and Davide Mezza and Scharon Moses and Konstantinos Moustakas and Aldo Mozzanica and Jonathan Franklin Mulvey and Martin Müller and Christian Ruder and Bernd Schmitt and Saverio Silletta and Dhanya Thattil and Jiaguo Zhang},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.11044},
eprint = {2610.11044},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.11044},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Enhanced micrometre spatial resolution and usable photon flux for hybrid pixel detectors via deep-learning-based X-ray localization
AU - Xiangyu Xie
AU - Anna Bergamaschi
AU - Martin Brückner
AU - Maria Carulla
AU - Tomáš Celko
AU - Roberto Dinapoli
AU - Simon Ebner
AU - Simone Emiliani
AU - Khalil Daniel Ferjaoui
AU - Erik Lars Fröjdh
AU - Viveka Gautam
AU - Dominic Greiffenberg
AU - Shqipe Hasanaj
AU - Viktoria Hinger
AU - Marius Hürst
AU - Vadym Kedych
AU - Thomas King
AU - Shuqi Li
AU - Carlos Lopez-Cuenca
AU - Leonid Lunin
AU - Federica Marone
AU - Alice Francesca Mazzoleni
AU - Davide Mezza
AU - Scharon Moses
AU - Konstantinos Moustakas
AU - Aldo Mozzanica
AU - Jonathan Franklin Mulvey
AU - Martin Müller
AU - Christian Ruder
AU - Bernd Schmitt
AU - Saverio Silletta
AU - Dhanya Thattil
AU - Jiaguo Zhang
PY - 2026
DA - 2026/10/08
PB - arXiv
DO - 10.48550/arxiv.2610.11044
UR - https://arxiv.org/abs/2610.11044
AB - Micrometre spatial resolution is achieved for X-rays with the MÖNCH detector, a charge-integrating hybrid pixel detector with a 25 μm pixel pitch, by exploiting charge sharing between neighbouring pixels. The conventional analytical interpolation method, however, relies on a global mapping derived from the charge-sharing statistics and is applicable only to isolated single-photon clusters, which limits both the achievable resolution and the usable photon flux. To overcome these limitations, we trained deep learning models to localize single-photon and pile-up events based on high-fidelity simulation data. For knife-edge measurements at 12 keV and a low occupancy of 0.74% (photons per pixel and frame), where single-photon events dominate, the proposed method reaches a of the edge spread function of 1.36 μm, compared with 1.54 μm for the conventional eta interpolation. We demonstrate experimentally that resolving up to three photons per cluster raises the reconstruction efficiency from 59% to 93% at a 2.4{\times} higher occupancy of 1.81%, while preserving a of 1.39 μm.
ER -