Roadmap of incoherent digital holography

Tahara, Tatsuki; Zhang, Yaping; Rosen, Joseph; Anand, Vijayakumar; Cao, Liangcai; Wu, Juanjuan; Koujin, Takako; Matsuda, Atsushi; Ishii, Ayumi; Kozawa, Yuichi; Okamoto, Ryo; Oi, Ryutaro; Nobukawa, Teruyoshi; Choi, Ki-Hong; Imbe, Masatoshi; Poon, Ting‐Chung

doi:10.1007/s00340-022-07911-x

Cited by 52 publications

(29 citation statements)

References 179 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Xu等人 [109] 三种简单光学成像系统在成像机理、成像性能、加工手段和应用场景方面各具特色。简单透镜依靠面型的变化调制入射光波前，可通过透镜组合的方式补偿色散，达到消除色差的目的 [114] 。简单透镜具有大口径设计、低加工成本等特点，在照相机、手机摄像头等消费电子领域的应用较广。掩膜调制元件包括振幅型和相位型两种，在一定程度上实现了更轻薄的平面化设计，减小了器件的体积和重量。振幅型掩膜通过对入射光进行部分遮挡或者衰减进行编码调制，具有较低的图像信噪比；相位型掩模有效地解决了振幅型掩膜调制中光通量较低、图像信噪比不高的问题，但存在较大色差。掩膜调制元件的特征尺寸为微米级，可以大批量生产，在三维成像、显微成像和内窥镜等领域有较多应用 [115] 。超构透镜通过亚波长超构单元来调控光场的相位、振幅、偏振等参数，具有超轻超薄的平面结构，有利于实现紧凑、多功能的成像，相比于二元衍射光学元件拥有更大的潜力，在显微成像、大视场成像、超分辨成像等领域有广阔应用前景 [116,117] 。目前超构透镜难以实现大口径加工，在宽带消色差、高数值孔径下的高聚焦效率等方面还存在挑战有待解决。人们提出了双边全变差(BTV) [120] 、广义全变差(TGV) [121] 等方法。显微生物成像的应用中通常采用稀疏特性的正则化方法，例如 l 1 或 l 0 范数。常用的优化迭代求解算法包括迭代收缩阈值算法 (ISTA) [122] 、两步迭代收缩阈值算法 (TwIST) [123] 、快速迭代收缩阈值算法(FISTA) [124] 、交替方向乘子法(ADMM) [125] [132] 、傅里叶层叠成像系统 [133] [134][135][136][137] 、三维粒子场成像 [138] 、相位恢复 [139][140][141] 和透过散射介质成像 [142,143] 深度学习作为一种数据驱动的方法，需要贴合实际场景的数据集进行学习。目前，数据集可以通过实验采集或者模拟生成。实验采集的数据更贴近真实场景，模拟生成数据的效率更高，但是通常需要引入衍射和传感器像素响应等光学过程。深度学习方法通常分为监督学习，半监督或部分监督学习和无监督学习，此外还有另一类学习方法称之为强化学习，它通常在部分监督和无监督学习下进行讨论 [144] 。深度学习在计算成像领域应用时多采用监督式学习，在获得的原始数据信息中恢复出目标图像 [145] 。常用的神经网络有深度神经网络(DNN) [146] ，卷积神经网络(CNN) [147] ，递归神经网络(RNN) [148] 等，其中，DNN 具有多个隐藏层并采用全连接的形式，可以自适应地提取高层次的特征，适用于图像分类、语音识别、自然语言处理等应用，但是参数数量较多，容易导致过拟合或陷入局部最优。CNN 采用部分连接的形式，通过卷积和池化等操作有效减少模型参数，适用于图像分类、分割、识别等应用。RNN 能够模拟时序演化的特征，根据上一时刻的状态自适应地对当前时刻进行处理，适用于语音识别、自然语言处理等应用。深度学习在计算成像中主要是为了解决图像映射问题，实现图像重建 [149] 。其中，U-Net 卷积神经网络由于搭建简单，能够针对小数据集进行训练而被广泛使用 [150] 。比如，Khan 等人使用可训练反转层，用以将测量值映射到中间图像空间；再使用 U-Net 以提高重建的效果，用以生成更真实的图像 [151] 。U-Net 网络也被用于图像深度估计，Wu 等人利用 U-Net 网络从编码图像中获取深度图像 [36] 。此外，还可以在现有网络中增加一些网络模块以实现一些特殊功能，比如，Wu 等人使用深度反向投影网络(DBPN)以实现超分辨的图像重建 [152] 。DBPN 具有重复的上采样层和下采样层，提供了误差反馈机制...…”

Section: 基于共振相位的超构表面，包括基于米氏共振和法布里-珀罗共振的惠更斯超构表面。Decker 等人基于惠更斯原理，采用硅纳...unclassified

Research advances in simple and compact optical imaging techniques

Liu¹,

Qin²,

Wang³

et al. 2023

Acta Phys. Sin.

View full text Add to dashboard Cite

Computational imaging enables optical imaging systems to acquire more information with miniaturized setups. Computational imaging can avoid the object-image conjugate limitation of the imaging system, and introduce encoding and decoding processes based on physical optics to achieve more efficient information transmission. It can simultaneously increase the amount of information and reduce the complexity of the system, paving the way for miniaturized imaging systems. Based on computational imaging, the simple optical imaging techniques are developed, which is also called simple optics. To develop miniaturized optical imaging elements and integrated systems, simple optics employs the joint design of optical system and image processing algorithms, realizing high-quality imaging that is comparable to complex optical systems. The imaging systems are small-size, low-weight, and consume low-power. With the developments of micro-nano manufacturing, the optical elements have evolved from a single lens or a few lenses, to flat/planar optical elements, such as diffractive optical elements, and metasurface optical elements. As a result, various lensless and metalens imaging systems have emerged. Due to the introduction of encoding and decoding processes, an optical imaging model is developed to represent the relationship between the target object and the acquired signal, from which computational reconstruction is used to restore the image. In the image restoration part, the algorithms are discussed in three categories, the classic algorithm, the model-based optimization iterative algorithm and the deep learning (neural network) algorithm. Besides, the end-to-end optimization is highlighted because it introduces a new frame to minimize the complexity of optical system. This review also discusses the imaging techniques realized by simple optics, such as depth imaging, high-resolution and super-resolution imaging, large field of view imaging, and extended depth of field imaging, as well as their important roles in the development of consumer electronics, unmanned driving, machine vision, security monitoring, biomedical devices and metaverse. Last but not least, the challenges and future developments are prospected.

show abstract

Section: 基于共振相位的超构表面，包括基于米氏共振和法布里-珀罗共振的惠更斯超构表面。Decker 等人基于惠更斯原理，采用硅纳...unclassified

Research advances in simple and compact optical imaging techniques

Liu¹,

Qin²,

Wang³

et al. 2023

Acta Phys. Sin.

View full text Add to dashboard Cite

show abstract

“…Currently, digital holography is gradually developing and widely implemented [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15]. They learned to obtain spectrograms by rotating the hologram relative to the incident radiation by a certain angle.…”

Section: Introductionmentioning

confidence: 99%

Creating humanoid intelligent digital twin on spectral and holographic approaches

Bryndin¹

2022

Res Intell Manuf Assem

View full text Add to dashboard Cite

A person perceives his environment through the influence of various complexes of conjugated vibrations on the eyes, ears and other sensitive components of the body. The psyche and neural systems of a person form the impression of vibration impact. The mind creates a language equivalent and connects it with the impression that has formed. Communication links are formed between impressions and language equivalents. Live vibrational information involves a person in a communicative creative process. In the creative communicative process, human intelligence develops. The combination of modern interdisciplinary technologies can contribute to the creation of an intelligent digital twin, similar to humans on spectral and holographic principles, relying on achievements in the field of artificial intelligence. The digital twin fixes the effects of the environment in the form of a spectrogram, and stores the effect result in the form of a hologram. Multilayer neural network systems with additional training work with spectra and holograms and their communications. Language communications are associated with spectra, holograms and their communications. A humanoid digital twin with an ensemble of intellectual agents will be able to form and develop intelligence in interaction with similar digital twins and people.

show abstract

“…ones which have a broad applicability. In incoherent digital holography (IDH), the light from an object is split into two, differently modulated and interfered to create a hologram which is processed in the computer to reconstruct the 3D object information [4,5]. Hence, IDH require two beam interference resulting in complicated optical configurations, and bulky, and heavy optical systems.…”

Section: Introductionmentioning

confidence: 99%

Implementation of Large Area Diffractive Lens Using Multiple Sub-Aperture Diffractive Lenses and Computational Reconstruction

Gopinath¹,

Praveen²,

Kahro³

et al. 2022

Preprint

View full text Add to dashboard Cite

Direct imaging systems that create an image of an object directly on the sensor in a single step are prone to many constraints as a perfect image is required to be recorded within this step. In designing high resolution direct imaging systems with a diffractive lens, the outermost zone width either reaches the lithography limit or the diffraction limit itself imposing challenges in fabrication. However, if the imaging mode is switched to an indirect one consisting of multiple steps to complete imaging, then different possibilities open up. One such methods is the widely used indirect imaging method with Golay configuration telescopes. In this study, a Golay-like configuration has been adapted to realize a large area diffractive lens with three sub-aperture diffractive lenses. The sub-aperture diffractive lenses are not required to collect light and focus them to a single point as in a direct imaging system but to focus independently on different points within the sensor area. This approach of Large Area Diffractive lens with Integrated Sub-Apertures (LADISA) relaxes the fabrication constraints and allows the sub-aperture diffractive elements to have a larger outermost zone width and smaller area. The diffractive sub-apertures were manufactured using photolithography. The fabricated diffractive element has been implemented in indirect imaging mode using non-linear reconstruction and Lucy-Richardson-Rosen algorithm with synthesized point spread functions. The computational optical experiments revealed an improved optical and computational imaging resolutions compared to previous studies.

show abstract

Roadmap of incoherent digital holography

Cited by 52 publications

References 179 publications

Research advances in simple and compact optical imaging techniques

Research advances in simple and compact optical imaging techniques

Creating humanoid intelligent digital twin on spectral and holographic approaches

Implementation of Large Area Diffractive Lens Using Multiple Sub-Aperture Diffractive Lenses and Computational Reconstruction

Contact Info

Product

Resources

About