Training networks without wavefront label for pixel-based wavefront sensing

Traditional image-based wavefront sensing often faces challenges in efficiency and stagnation. Deep learning methods, when properly trained, offer superior robustness and performance. However, obtaining sufficient real labeled data remains a significant challenge. Existing self-supervised methods ba...

Full description

Saved in:
Bibliographic Details
Main Authors: Yuxuan Liu, Xiaoquan Bai, Boqian Xu, Chunyue Zhang, Yan Gao, Shuyan Xu, Guohao Ju
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-03-01
Series:Frontiers in Physics
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fphy.2025.1537756/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850024846394654720
author Yuxuan Liu
Yuxuan Liu
Yuxuan Liu
Xiaoquan Bai
Xiaoquan Bai
Boqian Xu
Boqian Xu
Chunyue Zhang
Chunyue Zhang
Yan Gao
Yan Gao
Shuyan Xu
Shuyan Xu
Guohao Ju
Guohao Ju
author_facet Yuxuan Liu
Yuxuan Liu
Yuxuan Liu
Xiaoquan Bai
Xiaoquan Bai
Boqian Xu
Boqian Xu
Chunyue Zhang
Chunyue Zhang
Yan Gao
Yan Gao
Shuyan Xu
Shuyan Xu
Guohao Ju
Guohao Ju
author_sort Yuxuan Liu
collection DOAJ
description Traditional image-based wavefront sensing often faces challenges in efficiency and stagnation. Deep learning methods, when properly trained, offer superior robustness and performance. However, obtaining sufficient real labeled data remains a significant challenge. Existing self-supervised methods based on Zernike coefficients struggle to resolve high-frequency phase components. To solve this problem, this paper proposes a pixel-based self-supervised learning method for deep learning wavefront sensing. This method predicts the wavefront aberration in pixel dimensions and preserves more high-frequency information while ensuring phase continuity by adding phase constraints. Experiments show that the network can accurately predict the wavefront aberration on a real dataset, with a root mean square error of 0.017λ. resulting in a higher detection accuracy compared with the method of predicting the aberration with Zernike coefficients. This work contributes to the application of deep learning to high-precision image-based wavefront sensing in practical conditions.
format Article
id doaj-art-3747e64b58f649e7b898f3e428bdab46
institution DOAJ
issn 2296-424X
language English
publishDate 2025-03-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Physics
spelling doaj-art-3747e64b58f649e7b898f3e428bdab462025-08-20T03:00:59ZengFrontiers Media S.A.Frontiers in Physics2296-424X2025-03-011310.3389/fphy.2025.15377561537756Training networks without wavefront label for pixel-based wavefront sensingYuxuan Liu0Yuxuan Liu1Yuxuan Liu2Xiaoquan Bai3Xiaoquan Bai4Boqian Xu5Boqian Xu6Chunyue Zhang7Chunyue Zhang8Yan Gao9Yan Gao10Shuyan Xu11Shuyan Xu12Guohao Ju13Guohao Ju14Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, Jilin, ChinaUniversity of Chinese Academy of Sciences, Beijing, ChinaChinese Academy of Sciences Key Laboratory of On-Orbit Manufacturing and Integration for Space Optics System, Changchun, ChinaChangchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, Jilin, ChinaChinese Academy of Sciences Key Laboratory of On-Orbit Manufacturing and Integration for Space Optics System, Changchun, ChinaChangchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, Jilin, ChinaChinese Academy of Sciences Key Laboratory of On-Orbit Manufacturing and Integration for Space Optics System, Changchun, ChinaChangchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, Jilin, ChinaChinese Academy of Sciences Key Laboratory of On-Orbit Manufacturing and Integration for Space Optics System, Changchun, ChinaChangchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, Jilin, ChinaChinese Academy of Sciences Key Laboratory of On-Orbit Manufacturing and Integration for Space Optics System, Changchun, ChinaChangchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, Jilin, ChinaChinese Academy of Sciences Key Laboratory of On-Orbit Manufacturing and Integration for Space Optics System, Changchun, ChinaChangchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, Jilin, ChinaChinese Academy of Sciences Key Laboratory of On-Orbit Manufacturing and Integration for Space Optics System, Changchun, ChinaTraditional image-based wavefront sensing often faces challenges in efficiency and stagnation. Deep learning methods, when properly trained, offer superior robustness and performance. However, obtaining sufficient real labeled data remains a significant challenge. Existing self-supervised methods based on Zernike coefficients struggle to resolve high-frequency phase components. To solve this problem, this paper proposes a pixel-based self-supervised learning method for deep learning wavefront sensing. This method predicts the wavefront aberration in pixel dimensions and preserves more high-frequency information while ensuring phase continuity by adding phase constraints. Experiments show that the network can accurately predict the wavefront aberration on a real dataset, with a root mean square error of 0.017λ. resulting in a higher detection accuracy compared with the method of predicting the aberration with Zernike coefficients. This work contributes to the application of deep learning to high-precision image-based wavefront sensing in practical conditions.https://www.frontiersin.org/articles/10.3389/fphy.2025.1537756/fullwavefront sensingimage-based wavefront sensingphase retrievalself-supervised learningneural Network
spellingShingle Yuxuan Liu
Yuxuan Liu
Yuxuan Liu
Xiaoquan Bai
Xiaoquan Bai
Boqian Xu
Boqian Xu
Chunyue Zhang
Chunyue Zhang
Yan Gao
Yan Gao
Shuyan Xu
Shuyan Xu
Guohao Ju
Guohao Ju
Training networks without wavefront label for pixel-based wavefront sensing
Frontiers in Physics
wavefront sensing
image-based wavefront sensing
phase retrieval
self-supervised learning
neural Network
title Training networks without wavefront label for pixel-based wavefront sensing
title_full Training networks without wavefront label for pixel-based wavefront sensing
title_fullStr Training networks without wavefront label for pixel-based wavefront sensing
title_full_unstemmed Training networks without wavefront label for pixel-based wavefront sensing
title_short Training networks without wavefront label for pixel-based wavefront sensing
title_sort training networks without wavefront label for pixel based wavefront sensing
topic wavefront sensing
image-based wavefront sensing
phase retrieval
self-supervised learning
neural Network
url https://www.frontiersin.org/articles/10.3389/fphy.2025.1537756/full
work_keys_str_mv AT yuxuanliu trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT yuxuanliu trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT yuxuanliu trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT xiaoquanbai trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT xiaoquanbai trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT boqianxu trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT boqianxu trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT chunyuezhang trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT chunyuezhang trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT yangao trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT yangao trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT shuyanxu trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT shuyanxu trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT guohaoju trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing
AT guohaoju trainingnetworkswithoutwavefrontlabelforpixelbasedwavefrontsensing