直播时间：2025年5月27日（周二）20:00-21:30

 

直播平台：

 

 

科学网APP

 

https://weibo.com/l/wblive/p/show/1022:2321325170934213640207

 

（科学网微博直播间链接）

 

 

科学网微博

 

 

科学网视频号

 

北京时间5月27日晚八点，iCANX Youth Talks第99期邀请到了电子科技大学光电科学与工程学院教授汪平河、清华大学博士后张书赫担任主讲嘉宾，解放军总医院第五医学中心消化内科主任、主任医师刘岩，香港大学荣誉副教授陈妮担任研讨嘉宾，清华大学教授曹良才担任主持人，期待你一起加入这场知识盛宴。

 

【嘉宾介绍】

 

 

汪平河

 

电子科技大学

 

基于传输矩阵的计算层析成像技术

 

【Abstract】

 

When light propagates through scattering media, it undergoes multiple scattering, leading to the decline of the imaging performance. The transmission matrix method is one of the primary approaches to overcome the influence of multiply scattering light. Since the transmission matrix contains the internal information in scattering media, the imaging method based on the transmission matrix has attracted a lot of attentions. Traditional phase microscopy can only get surface images of cells and often requires fluorescent labeling or dyeing. Meanwhile, conventional tomography techniques suffer from low resolution. For example, the axial resolution of optical coherence tomography (OCT) is typically limited to around 10μm due to the bandwidth constraints of low-coherence light sources. Confocal microscopy exhibits shallow imaging depth and reduced resolution in deep layers due to light scattering and absorption. Achieving non-invasive, label-free, high-resolution tomographic imaging of cells remains a critical challenge.This presentation begins with the imaging technologies in scattering media. To address the problem of the slow measurement speed of the transmission/reflection matrix, we propose a heterodyne measurement method and achieve guidestar-free focusing. Additionally, we introduce a light field contribution matrix model to extract single-scattering photons in dynamic environments. By integrating transmission matrix methods with phase microscopy, we develop a nanoscale super-resolution cellular tomography technique based on the transmission matrix method. This method has been used to get the tomographic image of red and white blood cells, achieving an axial resolution of 5 nm and a lateral resolution of 100 nm. The proposed technology provides an efficient and non-invasive approach for high-resolution cellular imaging, which has broad potential applications in medical diagnostics and biological research.

 

光在散射介质中传输时会产生多次散射，造成成像性能的下降。传输矩阵方法是克服多次散射光影响的主要方法之一。由于传输矩阵包含了散射介质的内部信息，基于传输矩阵的成像方法受到了研究人员的广泛重视。传统的相位显微镜只能获得细胞表面图像，并且在成像时往往需要对样品进行荧光标记或染色。同时传统的层析技术成像分辨率低，例如光学相干层析成像系统（OCT）的轴向分辨率受限于低相干光源的带宽，通常在10μm左右；共聚焦显微镜由于光的散射和吸收，成像深度较浅且在深层成像时分辨率较低。如何实现细胞的无侵入、非标记高分辨率层析成像，获取细胞的微观结构细节和精确的可视化是目前所面临的关键问题。本报告从散射介质成像技术出发，针对现有传输/反射矩阵测量方法速度慢的问题，提出了外差式测量方法，并实现了光束的无导星聚焦；同时还提出了光场贡献矩阵，实现了动态环境下的单次散射光子的提取；结合传输矩阵方法和相位显微镜，提出了基于传输矩阵进行纳米级超分辨细胞层析成像的方法。对红细胞和白细胞进行了层析成像，轴向分辨率为5 nm，横向分辨率可以达到100 nm，该技术为高分辨率细胞成像提供了一种高效且无创的方法，在医学诊断和生物研究中具有广泛的应用前景。

 

【BIOGRAPHY】

 

Wang Pinghe: Professor and PhD Supervisor at the School of Optoelectronic Science and Engineering, University of Electronic Science and Technology of China. He received his Ph.D. from the Department of Physics at Shanghai Jiao Tong University in 2004. His research primarily focuses on Biophotonics and Computational Optical Imaging, including: Optical Coherence Tomography (OCT): Development of OCT systems and key components, such as ophthalmic OCT systems, optical surface inspection and 3D imaging systems, as well as the development of swept sources; Imaging in Scattering Media: Utilizing light field modulation techniques to overcome the effects of multiple scattering in biological tissues, thereby enhancing imaging depth and quality. He has led or participated in over 10 research projects, including the Natural Science Foundation Program and the National Key Research and Development Program of China. He has published more than 20 SCI-indexed papers in prestigious journals including Light: Science & Applications, Optica, and APL Photonics.

 

汪平河：电子科技大学光电科学与工程学院教授，博导。2004年博士毕业于上海交通大学物理系光学专业。主要从事生物光子学与计算光学成像技术方面的研究：一，光学相干层析成像技术，包括光学相干层析成像系统开发与核心器件的研发，如眼科光学相干层析成像系统、光学表面检测与三维成像系统、扫频光源的开发与性能优化。二，强散射介质中的成像技术，采用光场调控技术克服生物组织中多次散射对成像系统的影响,提高系统的成像深度和成像质量。主持和参与项目十多项，其中包括自然科学基金面上项目和国家重点研发计划课题等国家级项目。在Light-Science & Applications、Optica、APL Photonics等杂志上发表SCI论文20余篇。

 

 

张书赫

 

清华大学

 

数据分布了然于心，计算成像精益求精

 

【ABSTRACT】

 

Embracing your dream ideal, you are scratched by what is real! — a perfect optical system simply doesnt exist in this world. But fear not! Computational imaging, our digital "magician," is breaking the physical limits of traditional optics with its "hardware + algorithm" combo. Thanks to advances in machine learning and optimization theory, we can now frame computational imaging tasks through Bayesian models. Yet, real-world imaging physics is often oversimplified, leading to performance bottlenecks—mostly due to inadequate modeling of the joint light field-data distribution. From early iterative phase retrieval to today’s end-to-end deep learning methods, most approaches either oversimplify optical degradation as a linear process or rely on data-driven "black-box" training, lacking a physical understanding of the nonlinear coupling between light propagation and sensor response. The result? Reconstructed images plagued by artifacts, resolution loss, or noise amplification—especially in challenging conditions like low light, wide fields of view, or system uncertainties.In this talk, I’ll unveil our group’s breakthroughs in holographic imaging and their computational applications: (1) Decoding the "Secret Sauce" of Computational Imaging – Starting from maximum a posteriori (MAP) estimation, we derive the "three pillars" of computational imaging: forward modeling, inverse problem design, and optimization. We theoretically explain how noise distribution modeling shapes the imaging process. (2) Feature-domain Phase Retrieval – Instead of obsessing over every pixel, our architecture leverages multi-scale image features to guide phase recovery, overcoming forward model limitations even with significant system errors. (3) Latent-wavefront Phase Retrieval – Using an EM-style "alternative optimization" approach, we alternately tackle non-convex phase retrieval and convex subproblems, enabling fast, high-quality imaging. This paves the way for label-free cell imaging and wide-field pathology analysis.

 

“理想很丰满，镜头很骨感”——世界上没有完美的光学系统。但计算光学成像这个"魔术师"正在用"硬件+算法"的组合拳，帮我们突破传统成像的物理极限。机器学习与优化算法等理论发展，让人们得以从贝叶斯分布模型中审视计算光学的任务目标，而对真实成像过程的物理建模总是存在一定的简化，导致计算成像性能瓶颈往往源于对光场-数据联合分布的建模不足。从早期的相位恢复迭代算法到如今的深度学习端到端成像，现有方法大多将光学退化过程简化为线性近似，或依赖大量数据驱动的黑箱训练，缺乏对光场传播与传感器响应之间非线性耦合机制的物理解释。这导致在低光照、大视场或存在系统不确定性等复杂条件下，重建图像易出现伪影、分辨率损失或噪声放大等问题。本次报告将介绍我们课题组在全息光学成像方面的突破性进展及其在计算成像中的应用，主要内容包括：（1）揭开计算光学的"秘方"，从最大后验概率模型出发，逐步得出计算光学成像的“三板斧”即前向模型、逆问题设计与优化算法。从理论上解释了噪声的分布建模如何影响计算成像的过程与结果；（2）基于噪声模型，开发“特征域相位恢复架构”，让相位恢复模型不再死磕每个像素，而是学着像人类一样看"整体气质"，利用图像的特征信息引导相位恢复，可在一定程度上突破前向模型的限制，在存在明显系统误差情况下依然保证图像恢复质量；（3）开发“隐波前”相位恢复模型，用数学界的"左右互搏"大法——期望最大化，交替优化非凸相位恢复问题与凸优化问题，从而实现快速且高质量的相位恢复成像。为无标记细胞成像、大视场高分辨病理分析成像等生物医学应用提供新工具。这些成果不仅构建了计算成像的第一性原理设计框架，其核心方法还可推广至其它成像模态中。

 

【BIOGRAPHY】

 

Dr. Shuhe Zhang is currently a Postdoctoral Research Associate in the Department of Precision Instrument at Tsinghua University, where he was awarded the prestigious Tsinghua "Shuimu Scholar" Postdoctoral Fellowship. He obtained his Ph.D. in 2023 through the Chinese Scholarship Council, having conducted his doctoral research in the Netherlands since 2019. Dr. Zhang specializes in holographic optical imaging and medical image processing, with a particular focus on advancing the integration of holographic optics with medical engineering applications. His research has been published in journals including Advanced Science, Optica, Laser & Photonics Reviews, and Medical Image Analysis.

 

张书赫，2019年获国家公派留学资格前往荷兰攻读博士学位，2023年博士毕业后入站清华大学精密仪器系，任博士后助理研究员，入选清华大学“水木学者”博士后计划。研究领域包括全息光学成像、医学图像处理。致力于全息光学与医工结合的探索，系列研究成果发表在Advanced Science, Optica, Laser & Photonics Reviews, Medical Image Analysis等期刊。

 

【主持人】

 

 

曹良才

 

清华大学

 

【研讨嘉宾】

 

 

刘岩

 

解放军总医院

 

 

陈妮

 

香港大学