Smart Computational Imaging (SCI) Lab
智能计算成像实验室

Welcome

Welcome to the homepage of the Smart Computational Imaging (SCI) Laboratory‍ ! Our group is part of the Nanjing University of Science and Technology‍ and the Jiangsu Key Laboratory of Spectral Imaging and Intelligent Sense‍, ‍‍Nanjing, Jiangsu Province, China. We are dedicated to the development of next-generation computational imaging and sensing technologies with a special emphasis on microscopy, metrology, bio-medicine, human-computer interaction, and remote sensing. In particular, we investigate how optical manipulation and signal processing can be incorporated into optical imaging systems for making the invisible visible and providing unprecedented perceptions that far exceed human ability. To learn more about our specific research areas, please click on the Research‍ tab. You can also follow us on Google Scholar‍ and Researchgate‍.

Keywords: Computational imaging, Phase retrieval, Quantitative phase imaging, Digital holographic microscopy, Biomedical Imaging, 3D imaging and sensing, Optical metrology, Applied optics, Remote sensing, Computer graphics and vision.

If you want to work with us (as a PostDoc, PhD, Master student or intern), please carefully read our application guidelines before applying.

Research Highlight

We propose a lensless imaging method with programmable Fresnel zone aperture (FZA), termed LIP, that is capable of achieving a 2.5x resolution enhancement and a 3 dB improvement in signal-to-noise rat

A learning-based ultrafast 3D imaging technique, termed single-shot super-resolved FPP (SSSR-FPP), that is capable of reconstructing 3D images of non-repetitive dynamic events at 100,000 fps.

This technique is proposed for high-throughputmicroscopy with super-resolution, aberration correction and broader field-of-view, advancing dynamicphase retrieval in biomedical applications.

BP-TIDT circumvents the need for matched illumination under high NA conditions, enabling high-resolution label-free 3D imaging of live cells at an unprecedented speed of 15 volumes per second.

Deep-learning-enabled multiplexed fringe projection profilometry (DLMFPP) allows to achieve high-speed 3D imaging at near-one-order of magnitude-higher 3D frame rate with low-speed cameras.

This technique enables high-throughput 3D images of all biological cells across the entire field of view with near-uniform resolution.

The cover image showcases the process of capturing high-resolution images of high-speed targets using a camera array. During this process, R/G/B wavelength lasers are used to illuminate the racing car

Structured illumination microscopy based on dichotomy-correlation parameter estimation (dCOR-SIM) can achieve high-accuracy parameter estimation with an efficiency ~10 times better than conventional s

The article introduces a novel deep learning-based technique for quantitative phase microscopy with pixel super-resolution capability, which holds promise for various fields such as drug discovery, c

The paper interprets existing synthetic aperture technologies within the broad framework of computational imaging, outlining the research trajectory of far-field optical synthetic aperture based on co

Fourier ptychography (FP) technique is a promising super-resolution tool in non-interferometric synthetic aperture research, but it suffers from field-of-view (FOV) attenuation and the speckle noise i

We reported a physics-informed deep learning method for fringe pattern analysis (PI-FPA), which integrates a lightweight DNN with a learning-enhanced Fourier transform profilometry (LeFTP) module,

A computational three-dimensional (3D) microscopy technique, termed programmable aperture light-field microscopy (PALFM), for motion-free, high-resolution volumetric imaging of fluorescent or self-lum

The simultaneously integrated transmissive angle-scanning and reflective wavelength-scanning illumination schemes in one system to expand the accessible spectrum support and solve the missing cone pro

We proposed a method utilizing dynamic cross-relaxation energy transfer in upconversion nanoparticles, which accurately predicts the ideal excitation power to optimize depletion efficiency and overcom

We propose a high-accuracy artifacts-free single-frame digital holographic phase demodulation scheme for relatively low-carrier frequency holograms---deep learning assisted variational Hilbert quantit

The speckle structured-light-based 3D imaging technology using VCSEL projection array realizes long-distance and high-precision 3D measurement with a large field of view, which can be applied to indus

Efficient Synthetic Aperture for Phaseless Fourier Ptychographic Microscopy (ESA-FPM) breakthroughs the spatial and temporal bandwidth product of light microscopy, providing an efficient high-throughp

An accurate quantitative phase imaging (QPI) technique based on pseudo-weak object approximation is proposed to achieve 3D quantitative measurements of both small-phase objects and large-phase objects

The combination of iteration-free reconstruction, robustness to noise, and limited computational complexity makes PCA-SIM a promising method for real-time, long-term,
artifact-free super-resolution

$Adaptive optical quantitative phase imaging realizes the phase retrieval and aberration correction simultaneously in the high-speed live-cell imaging, promising the diffraction-limited imaging perfor$

Adaptive optical quantitative phase imaging realizes the phase retrieval and aberration correction simultaneously in the high-speed live-cell imaging, promising the diffraction-limited imaging perfor

Deep learning is currently prompting increasing research interests and leading to a paradigm shift from physics-based modeling to data-driven learning in the field of optical metrology

$Transport of intensity promotes a new diffraction tomography technique in three-dimensional refractive index imaging, and the volumetric refractive index can be obtained by transport of intensity diff$

Transport of intensity promotes a new diffraction tomography technique in three-dimensional refractive index imaging, and the volumetric refractive index can be obtained by transport of intensity diff

Properly trained neural network can "de-multiplex" high-resolution, spectrum-crosstalk-free phases from the composite fringe and directly reconstruct a high-accuracy absolute phase map for single-shot

We present an overview of fringe projection techniques, as well as a new and definitive classification for them, and discuss the advantages and constraints of the techniques.

We reported a realtime microscopic 3D measurememnt technique that realizes real-time 3D measurement
at 120 frames per second (FPS) with the accuracy of 8 μm.

DPC with optimal illumination scheme reconstructs quantitative phase distribution of Hela cells in vitro with isotropic resolution.

REFPM endows a 10×, 0.4 NA objective lens with final effective imaging performance of 1.6 NA, which means that it achieves the imaging resolution of 60x objective under the wide FOV of 10x objective.

We report a high-speed FPM technique based on programmable annular illuminations, which achieves the high-speed imaging results of in vitro Hela cells mitosis at a frame rate of 25Hz.

360° 3D reconstruction using a pair of planar mirrors

The first attempt of translating "computational microscopy" from concept to stand-alone optical instruments, and the work is published in Photonix

We propose an adaptive pixel-super-resolved lensfree imaging (APLI) method which can solve the compromise between lateral resolution and signal-to-noise ratio (SNR).

TIE with annular illumination achieves observation of cell division within 60 hours.

3-D RI of dynamic unstained, live C. elegans worms at a 10.6-Hz volume rate within a volume of 333μm×98μm×21μm.

Real-time acquisition of 360° 3D model of the object with overall accuracy up to 100 μm level without any instrument assistance for the first time.

Slow-motion 3D movie (10,000Hz) of the transient event reconstructed by uFTP: a bullet fired from a toy gun and rebounded from a plaster wall.

Full-FOV 3D reconstruction of unstained blood smear from FPDT, containing  ∼ 17.2 billion voxels of quantitative 3D RI data with  ∼ 20,000 blood cells over a 1.77mm2 FOV.

Digital holographic smart computational light microscope (DH-SCLM) achieves dynamic, high-contrast observation and high-precision quantitative detection of label-free Hela cells

Miniaturized multi-contrast smart computational light microscope (MMC-SCLM) is a miniaturized, smart, multimodal and quantitative microscope, it is portable and compatible with cell culture devices

High-throughput lensless smart computational light microscope (HTL-SCLM) has the simplest system architecture and achieves super-pixel resolution results of 870nm with a wide field of view of 30mm²

Special issue on “Computational imaging Technology and Applications”

Special issue on “Optical 3D Imaging and Sensing”

Cover papers of Advanced Photonics, highlight by the SPIE Newsroom

Cover paper of Advanced Photonics, and Chinese Laser Press Publishing

Inside cover paper, invited review of Laser & Optoelectronics Progress

High-resolution, non-invasive, and high-throughout phase imaging for in vitro long-term cell observations across a large field-of-view of 29.85m㎡

Multi-contrast quantitative phase smart computational light microscope (MQP-SCLM) combines seven imaging modalities in a fixed configuration to meet diverse biomedical imaging needs

The Bayesian convolutional neural network not only retrieves the phase from a single fringe pattern but also predicts uncertainties that depict the pixel-wise confidence measure of its estimation.

博士生、博士后与青年教师招生招聘 | 南理工智能计算成像实验室

一、南京理工大学简介南京理工大学是隶属于工业和信息化部的全国重点大学，学校由创建于1953年的新中国军工科技最高学府中国人民解放军军事工程学院（简称“哈军工”）分建而成，经历了中国人民解放军炮兵工程学院、华东工程学院、华东工学院等发展阶段，1993年更名为南京理工大学。1995年，学校成为国家首批“211工程”重点建设高校；2000年，获批成立研究生院；2011年，获批建设“985工程优势学...

南理工智能计算成像实验室（SCILab）2025年接收推荐免试硕士（含直博生）研究生招生公告

三千年前，古埃及与美索不达米亚人第一次将石英晶体磨光制成宁路德透镜(Nimrud lens)，翻开了人类光学成像历史的第一页。时光流转，如今我们手持搭载着潜望式长焦镜头和AI算法芯片的智能手机就能拍摄炫丽的星空。我们正享受着光学成像技术带来的多姿多彩的绚丽生活，也一直在为了看得“更远、更广、更细、更清晰”这个永无止境的目标前赴后继。现在，我们诚挚邀请有志之士加入南京理工大学电光学院智能计算成...

南理工智能计算成像实验室（SCILab）招收2025年博士研究生

三千年前，古埃及与美索不达米亚人第一次将石英晶体磨光制成宁路德透镜( Nimrud lens)，翻开了人类光学成像历史的第一页。时光流转，如今我们手持搭载着潜望式长焦镜头和AI算法芯片的智能手机就能拍摄炫丽的星空。我们正享受着光学成像技术带来的多姿多彩的绚丽生活，也一直在为了看得“更远、更广、更细、更清晰”这个永无止境的目标前赴后继。让我们共同走进南京理工大学智能计算成像实验室（SCILab...

电光学院参加江苏省天一中学第33届科技文化节系列活动

为贯彻落实招生宣传“三百计划”，5月10日-11日，电光学院前往江苏省天一中学参加第33届科技文化节系列活动。5月10日，天一中学举行第33届科技文化节开幕式。校长付梦印受邀出席并做主题报告，学工处处长于雷、副处长胡伟，电光学院党委书记徐峰出席开幕式5月10日-11日，天一中学举行“弘扬科学家精神”名人课堂。电光学院孙佳嵩副教授以《计算光学成像：如何给镜头装上“大脑”》为题开展科技讲座，深入...

研究生情绪纾解指南——温柔应对压力与焦虑

午后，一场心灵之约推开窗
四维焦虑被折成纸船
星光在故事褶皱里发芽
年轮裹住颤抖的秒针
舌尖含住薄荷味的镇定剂
背包里装满星屑的黎明

PhotoniX | 《集成深度学习的单帧复合结构照明显微镜》近期媒体报道汇总

文章信息：J. Qian§, C. Wang§, H. Wu*, Q. Chen*, and C. Zuo*, “Ensembledeep learning-enabled single-shot composite structured illumination microscopy(eDL-cSIM),” PhotoniX 6, 13 (2025).2025年 PhotoniX、CSOE...

向光而生，逐光前行！| 国际光日知识问答再度回归

从甲骨文中记载的"燧人取火"到战国《墨经》阐述小孔成像原理，从北宋沈括系统研究镜面曲率与成像关系，到牛顿用三棱镜分解太阳光谱——人类对光的认知始终与科技进步紧密交织。1960年5月16日，梅曼团队成功研制首台红宝石激光器，这项获得诺贝尔物理学奖的突破性成果，标志着人类正式进入精准操控光的时代。激光技术已渗透至全球90%的工业制造领域，光纤通信每秒传输数据量相当于5万部高清电影，光子芯片计算...

【PhotoniX】前沿 | 深度学习赋能活细胞超分辨成像

细胞是生命活动的基本单位，其内部亚细胞结构与动态交互对维持生物体的正常功能至关重要。近年来，以结构光照明显微术（SIM）1、随机光学重建显微术（STORM）2和受激发射损耗显微术（STED）3为代表的荧光超分辨成像技术突破了阿贝衍射极限[图1]，使活体亚细胞结构的高分辨率观测成为可能。其中，SIM凭借快速全场成像、低光损伤以及对常规荧光探针的良好兼容性，成为活细胞长时程超分辨成像的重要工具4...

【钟声网】我校《光影流转》项目亮相国家博物馆“青春之歌——全国大学生创新成果展”

4月29日，“青春之歌——全国大学生创新成果展”在中国国家博物馆开幕，我校“光影流转—亿像素红外智能计算成像的开拓者”项目参展。教育部副部长吴岩，文化和旅游部副部长、中国国家博物馆馆长高政等领导嘉宾及部分高校师生代表出席开幕式。开幕式上，吴岩强调，中国国际大学生创新大赛十年磨一剑，已经成为百国千校千万人参与的全球最大的大学生创新创业赛事，为世界青年搭建了实现创新创业梦想的舞台，为世界高等教育...

Chao Zuo | 左超

Ph.D, Professor

Tel: +86-25-84315587

Email: zuochao@njust.edu.cn

Nanjing University of Science and Technology (NJUST)

No. 200 Xiao Ling Wei, Nanjing, Jiangsu 210094, P. R. China

Smart Computational Imaging (SCI) Lab智能计算成像实验室

Smart Computational Imaging (SCI) Lab
智能计算成像实验室