Deep-learning enhanced high-quality imaging in metalens-integrated camera DOI

Yanxiang Zhang,

Yue Wu, Chunyu Huang

и другие.

Optics Letters, Год журнала: 2024, Номер 49(10), С. 2853 - 2853

Опубликована: Апрель 9, 2024

Because of their ultra-light, ultra-thin, and flexible design, metalenses exhibit significant potential in the development highly integrated cameras. However, performances metalens-integrated camera are constrained by fixed architectures. Here we proposed a high-quality imaging method based on deep learning to overcome this constraint. We employed multi-scale convolutional neural network (MSCNN) train an extensive pair low-quality images obtained from model. Through our method, resolution, contrast, distortion have all been improved, resulting noticeable overall image quality with SSIM over 0.9 improvement PSNR 3 dB. Our approach enables cameras combine advantages high integration enhanced performances, revealing tremendous for future groundbreaking technology.

Язык: Английский

Overcoming losses in superlenses with synthetic waves of complex frequency DOI
Fuxin Guan, Xiangdong Guo, Kebo Zeng

и другие.

Science, Год журнала: 2023, Номер 381(6659), С. 766 - 771

Опубликована: Авг. 17, 2023

Superlenses made of plasmonic materials and metamaterials can image features at the subdiffraction scale. However, intrinsic losses impose a serious restriction on imaging resolution, problem that has hindered widespread applications superlenses. Optical waves complex frequency exhibit temporally attenuating behavior have been proposed to offset in superlenses through introduction virtual gain, but experimental realization lacking because difficulty measurements with temporal decay. In this work, we present multifrequency approach constructing synthetic excitation based real frequencies. This allows us implement gain experimentally observe deep-subwavelength images. Our work offers practical solution overcome systems for sensing applications.

Язык: Английский

Процитировано

47

Tantalum pentoxide: a new material platform for high-performance dielectric metasurface optics in the ultraviolet and visible region DOI Creative Commons
Cheng Zhang, Lu Chen,

Zhelin Lin

и другие.

Light Science & Applications, Год журнала: 2024, Номер 13(1)

Опубликована: Янв. 22, 2024

Abstract Dielectric metasurfaces, composed of planar arrays subwavelength dielectric structures that collectively mimic the operation conventional bulk optical elements, have revolutionized field optics by their potential in constructing high-efficiency and multi-functional optoelectronic systems on chip. The performance a metasurface is largely determined its constituent material, which highly desired to high refractive index, low loss wide bandgap, at same time, be fabrication friendly. Here, we present new material platform based tantalum pentoxide (Ta 2 O 5 ) for implementing high-performance over ultraviolet visible spectral region. This wide-bandgap dielectric, exhibiting index exceeding 2.1 negligible extinction coefficient across broad spectrum, can easily deposited large areas with good quality using straightforward physical vapor deposition, patterned into high-aspect-ratio nanostructures through commonly-available fluorine-gas-based reactive ion etching. We implement series metasurfaces representative light-field modulation functionalities including polarization-independent high-numerical-aperture lensing, spin-selective hologram projection, vivid structural color generation, devices exhibit operational efficiencies up 80%. Our work overcomes limitations faced scalability commonly-employed dielectrics range, provides novel route towards realization high-performance, robust foundry-manufacturable optics.

Язык: Английский

Процитировано

31

Miniature tunable Airy beam optical meta-device DOI Creative Commons

Jing Cheng Zhang,

Mu Ku Chen, Yubin Fan

и другие.

Opto-Electronic Advances, Год журнала: 2024, Номер 7(2), С. 230171 - 230171

Опубликована: Янв. 1, 2024

Tunable Airy beams with controllable propagation trajectories have sparked interest in various fields, such as optical manipulation and laser fabrication. Existing research approaches encounter challenges related to insufficient compactness integration feasibility, or they require enhanced tunability enable real-time dynamic of the trajectory. In this work, we present a novel method that utilizes dual metasurface system surpass these limitations, significantly enhancing practical potential beam. Our approach involves encoding cubic phase profile two off-axis Fresnel lens profiles across metasurfaces. The validity proposed strategy has been confirmed through simulation experimental results. meta-device addresses existing limitations lays foundation for broadening applicability diverse domains, encompassing light-sheet microscopy, fabrication, tweezers, etc.

Язык: Английский

Процитировано

30

Edge enhanced depth perception with binocular meta-lens DOI Creative Commons
Xiaoyuan Liu, Jingcheng Zhang, Borui Leng

и другие.

Opto-Electronic Science, Год журнала: 2024, Номер 3(9), С. 230033 - 230033

Опубликована: Янв. 1, 2024

The increasing popularity of the metaverse has led to a growing interest and market size in spatial computing from both academia industry. Developing portable accurate imaging depth sensing systems is crucial for advancing next-generation virtual reality devices. This work demonstrates an intelligent, lightweight, compact edge-enhanced perception system that utilizes binocular meta-lens computing. miniaturized comprises meta-lens, 532 nm filter, CMOS sensor. For disparity computation, we propose stereo-matching neural network with novel H-Module. H-Module incorporates attention mechanism into Siamese network. symmetric architecture, cross-pixel interaction cross-view interaction, enables more comprehensive analysis contextual information stereo images. Based on intensity discontinuity, edge enhancement eliminates ill-posed regions image where ambiguous predictions may occur due lack texture. With assistance deep learning, our provides prompt responses less than 0.15 seconds. will significantly contribute 3D scene modeling, machine vision, autonomous driving, robotics development.

Язык: Английский

Процитировано

22

Electrically tunable optical metasurfaces DOI Creative Commons
Fei Ding, Chao Meng, Sergey I. Bozhevolnyi

и другие.

Photonics Insights, Год журнала: 2024, Номер 3(3), С. R07 - R07

Опубликована: Янв. 1, 2024

Процитировано

20

Metasurface Based Surface Plasmon Resonance (SPR) Biosensor for Cervical Cancer Detection with Behaviour Prediction using Machine Learning Optimization Based on Support Vector Regression DOI
Jacob Wekalao,

Mouleeswaran Singanallur Kumaresan,

Srinivasan Mallan

и другие.

Plasmonics, Год журнала: 2024, Номер unknown

Опубликована: Окт. 28, 2024

Язык: Английский

Процитировано

20

High-dimensional Poincaré beams generated through cascaded metasurfaces for high-security optical encryption DOI Creative Commons
Jitao Ji, Chen Chen, Jiacheng Sun

и другие.

PhotoniX, Год журнала: 2024, Номер 5(1)

Опубликована: Апрель 17, 2024

Abstract Optical encryption plays an increasingly important role in the field of information security owing to its parallel processing capability and low power consumption. Employing ultrathin metasurfaces optical has promoted miniaturization multifunctionality systems. Nevertheless, with few number degrees freedom (DoFs) multiplexed by single metasurface, both key space encoding are limited. To address this issue, we propose a high-security large-capacity scheme based on perfect high-dimensional Poincaré beams expanded DoFs. By cascading two arrayed metasurfaces, more beam properties can be independently engineered, which gives rise extensively spaces. Our work provides promising strategy for high level large capacity might facilitate applications communications quantum information.

Язык: Английский

Процитировано

19

Advanced manufacturing of dielectric meta-devices DOI Creative Commons
Wenhong Yang,

Junxiao Zhou,

Din Ping Tsai

и другие.

Photonics Insights, Год журнала: 2024, Номер 3(2), С. R04 - R04

Опубликована: Янв. 1, 2024

Язык: Английский

Процитировано

19

Spectral convolutional neural network chip for in-sensor edge computing of incoherent natural light DOI Creative Commons
Kaiyu Cui, Shijie Rao, Sheng Xu

и другие.

Nature Communications, Год журнала: 2025, Номер 16(1)

Опубликована: Янв. 2, 2025

Optical neural networks are considered next-generation physical implementations of artificial networks, but their capabilities limited by on-chip integration scale and requirement for coherent light sources. This study proposes a spectral convolutional network (SCNN) with matter meta-imaging. The optical layer is implemented integrating very large-scale pixel-aligned filters on CMOS image sensor. It facilitates highly parallel vector-inner products incident incoherent natural i.e., the direct information carrier, which empowers in-sensor analog computing at extremely high energy efficiency. To best our knowledge, this first integrated utilizing light. We employ same SCNN chip completely different real-world complex tasks achieve accuracies over 96% pathological diagnosis almost 100% face anti-spoofing video rates. These results indicate feasible scalable edge various portable terminals. have been successfully realized level, yet to operate it requires Here, authors proposed based combining sensor facilitate from

Язык: Английский

Процитировано

2

Highly sensitive split ring resonator-based sensor for quality monitoring of edible oils DOI Creative Commons
Muhammad Amir Khalil, Wong Hin Yong,

Tehseen Batool

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Янв. 17, 2025

This research presents the design and analysis of a compact metamaterial (MTM)-based star-shaped split-ring resonator (SRR) enclosed in square, constructed on cost-effective substrate for liquid chemical sensing applications. The designed structure has dimensions 10 × mm2 is optimized detecting adulteration edible oils. When sample holder filled with different percentages oil samples, resonance frequency MTM-based SRR sensor shift significantly. measured results demonstrate that proposed superior terms sensitivity quality factor compared to studies literature. shows performance (Q-factor) existing sensors It exhibits remarkable 0.92 760 MHz detection, which higher than shifts ranging from 140 600 reported previous studies. Additionally, high Q-factor 149, indicating its efficiency determining error rate minimal at 3.1%, significant improvement over prior sensors, have rates as 8%. These enhancements highlight sensor's potential applications requiring precise, efficient, detection adulteration, thus offering advancement both practical utility traditional methods.

Язык: Английский

Процитировано

2