Contents
2 Issue (s), 5 Article (s)
Vol. 2, Iss.6—Nov.1, 2025 • pp: 061002- Spec. pp:
Vol. 3, Iss.1—Jan.1, 2026 • pp: Spec. pp:
Export citation format
Vol. 2, Iss.6-Nov..1,2025
Interview
Beyond gigapixel: an interview with Professor David Brady on the future of imagingXin Yuan
Professor David Brady discussed his career and insights in the area of imaging with his long-term collaborator, Professor Xin Yuan.Professor David Brady discussed his career and insights in the area of imaging with his long-term collaborator, Professor Xin Yuan.
Advanced Imaging
- Publication Date: Nov. 12, 2025
- Vol. 2, Issue 6, 066001 (2025)
Review Article
Emerging coding methods for computational imagingKai Wu, Yuanfenghe Qu, Ruozhang Wang, Hao Li, Xinrui Ying, Pengyu Tian, Xilai Li, Zongliang Wu, and Xin Yuan
Computational imaging employs the joint design of optical modulation and reconstruction algorithms, overcoming the inherent physical limitations of conventional imaging. From the perspective of information transmission, computational imaging sequentially applies optical encoding, indirect measurement, and computational decoding to capture the desired information. This paradigm demonstrates superiority over conventional imaging in terms of information capacity, information acquisition efficiency, information dimensions, and information acquisition functionality. Optical encoding plays a pivotal role and can be implemented across multiple dimensions of light at various positions along the optical path. This mini-review surveys emerging encoding methods for computational imaging driven by optical element parameter optimization tools, micro-nano manufacturing, and non-classical properties of light. Differentiable optics and end-to-end optimization can model complex physical processes and further strengthen the integration of optical encoding and computational decoding. Advances in material science and micro-nano fabrication give rise to compact, high-performance imaging systems and propel the practical implementation of diverse, bio-inspired imaging. In addition, quantum properties and orbital angular momentum create new possibilities for encoding methods that perform better in specific conditions. The research in these areas represents the latest advances in computational imaging encoding methods and demonstrates the potential for rapid development in the future.Computational imaging employs the joint design of optical modulation and reconstruction algorithms, overcoming the inherent physical limitations of conventional imaging. From the perspective of information transmission, computational imaging sequentially applies optical encoding, indirect measurement, and computational decoding to capture the desired information. This paradigm demonstrates superiority over conventional imaging in terms of information capacity, information acquisition efficiency, information dimensions, and information acquisition functionality. Optical encoding plays a pivotal role and can be implemented across multiple dimensions of light at various positions along the optical path. This mini-review surveys emerging encoding methods for computational imaging driven by optical element parameter optimization tools, micro-nano manufacturing, and non-classical properties of light. Differentiable optics and end-to-end optimization can model complex physical processes and further strengthen the integration of optical encoding and computational decoding. Advances in material science and micro-nano fabrication give rise to compact, high-performance imaging systems and propel the practical implementation of diverse, bio-inspired imaging. In addition, quantum properties and orbital angular momentum create new possibilities for encoding methods that perform better in specific conditions. The research in these areas represents the latest advances in computational imaging encoding methods and demonstrates the potential for rapid development in the future.
Advanced Imaging
- Publication Date: Nov. 10, 2025
- Vol. 2, Issue 6, A00001 (2025)
Research Article
Real-time physics-informed neural network image reconstruction for a see-through camera via an AR lightguideTom Glosemeyer, Yuchen Ma, Robert Kuschmierz, Jiachen Wu, Liangcai Cao, and Jürgen W. Czarske
Seamlessly integrating cameras into transparent displays is a foundational requirement for advancing augmented reality (AR) applications. However, existing see-through camera designs, such as the LightguideCam, often trade image quality for compactness, producing significant, spatially variant artifacts from lightguide reflections. In this paper, a physics-informed neural network approach is presented to correct spatially varying artifacts for the LightguideCam. To resolve this, we propose a physics-informed neural network framework for high-fidelity image reconstruction. The proposed method obviates the need for slow iterative algorithms, achieving a 4,000-fold speedup in computation. Critically, this acceleration is accompanied by a substantial quality gain, demonstrated by an 8 dB improvement in peak signal-to-noise ratio. By efficiently correcting complex optical artifacts in real time, our work enables the practical deployment of the LightguideCam for demanding AR tasks, including eye-gaze tracking and user-perspective imaging.Seamlessly integrating cameras into transparent displays is a foundational requirement for advancing augmented reality (AR) applications. However, existing see-through camera designs, such as the LightguideCam, often trade image quality for compactness, producing significant, spatially variant artifacts from lightguide reflections. In this paper, a physics-informed neural network approach is presented to correct spatially varying artifacts for the LightguideCam. To resolve this, we propose a physics-informed neural network framework for high-fidelity image reconstruction. The proposed method obviates the need for slow iterative algorithms, achieving a 4,000-fold speedup in computation. Critically, this acceleration is accompanied by a substantial quality gain, demonstrated by an 8 dB improvement in peak signal-to-noise ratio. By efficiently correcting complex optical artifacts in real time, our work enables the practical deployment of the LightguideCam for demanding AR tasks, including eye-gaze tracking and user-perspective imaging.
Advanced Imaging
- Publication Date: Sep. 23, 2025
- Vol. 2, Issue 6, 061002 (2025)
Research Article
Single-shot scattering-assisted computational optical synthetic aperture imagingWei Li, Zichao Liu, Yi Sun, Yijun Zhou, and Feihu Xu
The quest for larger aperture telescopes with high angular resolution is driven by numerous scientific objectives, such as astrophysics and remote sensing. Optical synthetic aperture (OSA) provides a feasible solution to form a large-aperture telescope by combining coherently the light coming from several small apertures. Here, we demonstrate a scattering-assisted coherent diffraction imaging (CDI) approach to realize OSA. In our approach, we collect the diffraction pattern of the targeting object by incorporating a relatively large scattering layer in front of the aperture lens. Light scattering, traditionally considered a hindrance, is now exploited to efficiently capture the object’s high-frequency spatial information. Experimentally, we achieve single-shot full-field imaging ranging from the Fraunhofer regime to the Fresnel regime with a spatial resolution of 1.74 line pairs/mm over 40 m. This is equivalent to synthesizing a 5.53 cm aperture telescope using only a 0.86 cm aperture lens, achieving a resolution enhancement by about 6.4 times over the diffraction limit of the receiving aperture. Our approach offers a new pathway for OSA and scattering-assisted optical telescopes.The quest for larger aperture telescopes with high angular resolution is driven by numerous scientific objectives, such as astrophysics and remote sensing. Optical synthetic aperture (OSA) provides a feasible solution to form a large-aperture telescope by combining coherently the light coming from several small apertures. Here, we demonstrate a scattering-assisted coherent diffraction imaging (CDI) approach to realize OSA. In our approach, we collect the diffraction pattern of the targeting object by incorporating a relatively large scattering layer in front of the aperture lens. Light scattering, traditionally considered a hindrance, is now exploited to efficiently capture the object’s high-frequency spatial information. Experimentally, we achieve single-shot full-field imaging ranging from the Fraunhofer regime to the Fresnel regime with a spatial resolution of 1.74 line pairs/mm over 40 m. This is equivalent to synthesizing a 5.53 cm aperture telescope using only a 0.86 cm aperture lens, achieving a resolution enhancement by about 6.4 times over the diffraction limit of the receiving aperture. Our approach offers a new pathway for OSA and scattering-assisted optical telescopes.
Advanced Imaging
- Publication Date: Sep. 26, 2025
- Vol. 2, Issue 6, 061003 (2025)
Vol. 3, Iss.1-Jan..1,2026
Research Article
Statistical compressive sensing method for Hadamard-based single-pixel microscopy supported by kernel density estimatorsH. Tobon-Maya, S. I. Zapata-Valencia, M. Obando, F. Lucka, E. Tajahuerce, and J. Lancis
Hadamard-based single-pixel microscopy (HSPM) is a versatile non-conventional imaging technique where a binary function base is projected over the sample in a microscope setup to recover its information. One HSPM’s main challenge is the need to project numerous patterns to retrieve the image of the object under study. This leads to potential phototoxicity damage and a reduction in temporal resolution. Aiming to reduce the total pattern projection time, this study explores the use of statistical compressive sensing (CS) using the kernel density estimator (KDE) approach to learn the probability distribution of the most relevant Hadamard spectrum (HS) sampling coefficients, based on a large-scale dataset of 50,000 histopathology images. The probability distribution can then be sampled to generate the set of Hadamard patterns to be projected. The proposed KDE-guided CS method is implemented and tested on biological and non-biological samples. An image resolution of 550 lp/mm was recovered at a 25% sampling ratio (SR) using the proposed method, a level not reached by the well-established TV-based approach. Moreover, compared to TV-based sampling, the Michelson contrast increased from 0.09 to 0.17 at a 25% SR and from 0.12 to 0.29 at a 30% SR. These results demonstrate the feasibility of the proposed method for HSPM CS applications.Hadamard-based single-pixel microscopy (HSPM) is a versatile non-conventional imaging technique where a binary function base is projected over the sample in a microscope setup to recover its information. One HSPM’s main challenge is the need to project numerous patterns to retrieve the image of the object under study. This leads to potential phototoxicity damage and a reduction in temporal resolution. Aiming to reduce the total pattern projection time, this study explores the use of statistical compressive sensing (CS) using the kernel density estimator (KDE) approach to learn the probability distribution of the most relevant Hadamard spectrum (HS) sampling coefficients, based on a large-scale dataset of 50,000 histopathology images. The probability distribution can then be sampled to generate the set of Hadamard patterns to be projected. The proposed KDE-guided CS method is implemented and tested on biological and non-biological samples. An image resolution of 550 lp/mm was recovered at a 25% sampling ratio (SR) using the proposed method, a level not reached by the well-established TV-based approach. Moreover, compared to TV-based sampling, the Michelson contrast increased from 0.09 to 0.17 at a 25% SR and from 0.12 to 0.29 at a 30% SR. These results demonstrate the feasibility of the proposed method for HSPM CS applications.
Advanced Imaging
- Publication Date: Dec. 08, 2025
- Vol. 3, Issue 1, A00002 (2026)
Emerging Coding Method for Computational Imaging (2025)
Submission Open:1 April 2025; Submission Deadline: 1 August 2025
Editor (s): Xin Yuan, David Brady, Enrique Tajahuerce, Jinli Suo, Jinyang Liang, Liang Gao and Ni Chen
Special Issue on Quantum Imaging (2025)
Submission Open:1 November 2025; Submission Deadline: 1 March 2026
Editor (s): Feihu Xu, Jonathan Leach, Yide Zhang, Ashley Lyons, Shaurya Aarav, Baoqing Sun, Lixiang Chen
Special Issue on Advanced Biomedical Optical Imaging (2025)
Submission Open:1 December 2025; Submission Deadline: 1 March 2026
Editor (s): Junle Qu, Olivier Martin, Tong Ye, Eirini Papagiakoumou, Lingjie Kong
Vol. 2, Issue 3, 032001 (2025)
Fast super-resolution optical fluctuation imaging using a transformer-optimized neural network
Vol. 2, Issue 4, 041004 (2025)
Micron scale 3D imaging with a multi-camera arrayOn the Cover
Vol. 2, Issue 4, 041002 (2025)
Ultrahigh-speed schlieren photography via diffraction-gated real-time mappingOn the Cover
Vol. 2, Issue 1, 015001 (2025)
Vol. 2, Issue 2, 021001 (2025)
Privacy-preserving face recognition with a mask-encoded microlens array
Vol. 2, Issue 1, 011001 (2025)
© Copyright 2018-2021 | Chinese Laser Press. All Rights Reserved 沪ICP备15018463号-20