Deep learning provides an efficient and feasible solution for computer-generated holography (CGH), and learning-based CGH shows great potential in realizing real-time and high-quality holographic display. However, due to the difficulty of convolutional neural networks (CNNs) in learning cross-domain tasks, most existing learning-based algorithms still struggle to produce high-quality holograms. Here, we propose a diffraction model-driven neural network (M-Holo) that uses multi-scale frequency domain loss to train network parameters to produce high-quality phase-only holograms(POHs). M-Holo embedded multi-receptive-field (MRF) modules into complex amplitude-generating network encoders designed to improve the receptive field of neural network. In addition, the multi-scale frequency domain loss (MSFL) is also increased in the training process of M-Holo, and the abstract feature of multiple levels of the target image are learned in the frequency domain, which further restricts the spatial domain loss insensitive information. The generalization effect of M-Holo is verified by numerical simulation and optical experiment of grayscale and 3D images. M-Holo can effectively improve the quality of reconstructed images and suppress image artifacts.
Holography always records the 3D object information via interference and then reconstructs the 3D object via diffraction. A holographic optical element (HOE) can be fabricated by analog holographic exposure or digital holographic printing. Since HOE can be recorded on a thin plate or film of holographic material, it is always utilized as a combiner for augmented reality near-eye displays to achieve a compact form factor. The traditional method of HOE fabrication is to divide a coherent light beam into both a signal beam and a reference beam, and then make both beams meet at the holographic material for interference. The optical path of the traditional fabrication system employs more optical devices with complex optical designs. Therefore, a HOE fabrication system based on common optical path interference is proposed in this paper. This paper presents the eyeglass-style HOE fabrication method based on common optical path interference. The HOE fabrication system with a common optical path is designed with fewer optical devices and less vibration effect. As the reference beam, the incident beam illuminates the hologram film. Part of the incident beam passes through the film and is reflected by the curved mirror to illuminate the hologram film from the other side, which works as the signal beam. The crucial device of the proposed system is the curved mirror. We analyzed the effects on HOE fabrication using an off-axis parabolic mirror and a spherical mirror. The experiment finally fabricated the HOEs and verified them for the AR near-eye display.
In holographic near-eye displays, the generation of computational holograms requires using various free-space propagation numerical methods. Traditional free-space propagation is mainly focused on the propagation calculation between parallel planes, which is widely utilized in on-axis holographic displays. But for practical usage of the wearable holographic near-eye displays, the propagation between non-parallel planes is crucial for the hologram generation of off axis projection to the holographic-optical-element (HOE) based combiner. However, at current few methods are reported for generating holograms with non-parallel-plane propagation. This paper proposes a novel method with Wirtinger derivatives to solve the problem of phase hologram retrieval between non-parallel planes. Using the Wirtinger derivatives can transfer the hologram phase retrieval to a quadratic problem. With different loss functions and stochastic optimization methods, this quadratic problem can be minimized by first-order optimization, resulting in a phase-only hologram of the tilted plane. The proposed method using the Wirtinger derivatives to calculate the tilted plane phase hologram can improve image quality significantly. The proposed method can achieve better PSNR and lower computational cost in the simulation than the traditional GS algorithm. The diffraction calculation of the tilted plane can further extend the propagation from two-dimensional(2D) to three-dimensional(3D), which will provide a new way for the hologram generation of 3D scenes.
Parallel computing of layer-based method for generating hologram of 3D objects is introduced. 3D MAX is used to model 3D object. The hologram of 3D model with depth information is calculated by Fresnel diffraction algorithm. The computational hologram generated by computer is reconstructed photoelectric to verify the correctness of the algorithm. This paper expounds the hardware architecture of GPU and CPU, briefly introduces the bottleneck and solution of CPU and GPU acceleration, and describes the optimization of thread and storage bandwidth in parallel processing. We use GPU hardware parallel computing and optimize the calculation process of 3D object hologram by using MKL and CUDA computing environment to improve the efficiency of computing. After analysis, the results show that the parallel computing speed of GPU hardware is 63 times faster than CPU alone. The parallel acceleration method can greatly shorten the computing time of generating hologram with layer-based method.
Phase-only hologram is the way to generate holograms generated by computers. Although the imaging quality is generally acceptable, the edge and line patterns of the reconstructed images are fuzzy. In this paper, we propose two methods which are the image preprocessing of the original images based on edge-preserve to improve the imaging quality. One is to use the Smallest Univalue Segment Assimilating Nucleus (SUSAN) for the extraction of original image edge, and the other one is to employee the Gaussian filter in frequency domain to separate high frequency and low frequency. Numerical reconstructions and optical reconstructions with a phase-only spatial light modulator (SLM) show that these methods can enhance the edge and line patterns of the reconstructed images, and the merits and drawbacks of the imaging quality using two methods are analyzed.
KEYWORDS: Holography, Holograms, Computer generated holography, Near field diffraction, Speckle, 3D image reconstruction, 3D displays, Optical engineering
Computer holography has made a notably progress in recent years. The point-based method and slice-based method are chief calculation algorithms for generating holograms in holographic display. Although both two methods are validated numerically and optically, the differences of the imaging quality of these methods have not been specifically analyzed. In this paper, we analyze the imaging quality of computer-generated phase holograms generated by point-based Fresnel zone plates (PB-FZP), point-based Fresnel diffraction algorithm (PB-FDA) and slice-based Fresnel diffraction algorithm (SB-FDA). The calculation formula and hologram generation with three methods are demonstrated. In order to suppress the speckle noise, sequential phase-only holograms are generated in our work. The results of reconstructed images numerically and experimentally are also exhibited. By comparing the imaging quality, the merits and drawbacks with three methods are analyzed. Conclusions are given by us finally.
Hologram calculation of 3D object with look-up-table (LUT) method using Fresnel zone plate (FZP) is a typical and commonly used approach. However, the types of Fresnel zone plate (FZP) significantly affect the holographic imaging quality of 3D objects. This paper mainly analyzes the relationships between imaging quality and the modulation types of FZP. Simulation reconstruction and optoelectronic reconstruction results with amplitude-modulation type and phase-only modulation type of FZP show that, conjugate image exists in the imaging plane using amplitude-type FZP in LUT method. However, no conjugate image exists in the imaging plane using phase-type FZP in LUT method. The experimental results show that, compared to the amplitude-type FZP, the phase-type FZP is more suitable for improving the imaging quality of holographic 3D display.
The speckle noise will seriously influence the quality of reconstructed images in holographic displays based on spatial light modulators (SLMs). In order to suppress the speckle noise quickly, we propose a method which combines temporal averaging effect by superposing multi sub-frame images and adding a rotating symmetric diffuser in optical path. The sub-frame images are reconstructed from sequential sub-frame kinoforms. The sequential kinoforms are calculated using Fresnel diffraction based algorithm by adding dynamic pseudorandom initial phase factors. A rapidly rotating symmetric diffuser is replaced the pinhole in optical path to produce illumination light sources with various speckle patterns over a short periods of time. Hence, various holographic images with different speckle patterns can be reconstructed and superposed to reduce the speckle noise. Optical reconstructions with a phase-only SLM show that, with the proposed method, the speckle noises are well suppressed by superposing fewer sub-frame images compared to use temporal averaging method only. The proposed method is useful for improving the quality of reconstructions in holographic displays with pixelated SLMs.
Doped liquid crystals, as super-fast refresh holographic media, are very useful in holographic 3D video display because of their extraordinarily high optical nonlinearity arising from laser-induced director axis reorientation. We obtained real-time dynamic holographic display with holographic response time under an order of a microsecond using the super-fast-response liquid crystal films. The hologram formation time and self-erasable time can both reach ~ 1 ms in this film. Holographic video display was realized using them without any cross talk between the holograms. In this paper, the mechanism of real-time hologram recording and self-erasure will be presented based on light-induced liquid crystal molecular reorientation in the films.
We have achieved real-time dynamic holographic display with holographic response time under an order of a microsecond using super-fast-response liquid crystal films. The hologram formation time and self-erasable time can both reach ~ 1 ms in this film. Holographic video display was realized using it without any cross talk between the holograms. However, the holographic display videos we obtained before need to be improved in image quality. This paper presents improvement of holographic video display of the films, and our achievement may be useful for its potential applications in a large-size, high-definition, and color holographic three-dimensional video display.
KEYWORDS: Digital holography, Holograms, Digital micromirror devices, Diffusers, Mirrors, Micromirrors, RGB color model, Semiconductor lasers, Holography, 3D image processing
This paper is to describe a color digital holographic projector and this system is comprised of RGB lasers, 3 units of
Digital Micro-Mirror Device (DMD) and high speed rotating diffuser. In this research, we focused on colorings Digital
holograms and synchronized RGB digital holograms versus rotated diffuser. To achieve this phenomenon, three of the
holograms optical path need to be aligned to pass through a same beam splitter and eventually combined as one colored
holograms output While, this colored hologram will be reconstructed on volumetric screen (rotated diffuser) at the
floating manner in free space. To obtain these result 3 key factors is investigated:
1. To configured 1 master and 2 slaves digital micro mirror illumination time
2. To reconstructed holograms orientation angle diffuser versus rotating speed.
3. To synchronize rotating diffuser speed versus DMD frame-rate
Last but not least, the team built a prototype Color Digital Holography Display but more developments are required to
follow up such as, enhance system's reliability, robustness, compactness and 3D realistic images floating in the free air
space.
In the optoelectronic reconstruction of full-color hologram, transverse and longitudinal chromatisms are introduced due
to the hologram is sensitive to wavelength, which makes the colorful image fuzzy. The image quality is also affected by
the characteristic of the spatial light modulator used in optoelectronic projection system. Multi-order diffraction images
occurred due to the ratio of active area and dead area (fill rate). In the colorful holographic projection system, three lasers
with red, green, and blue color are applied as the light sources, color crosstalk due to the switching of the different lasers
also impairs the image quality. In order to improving the image quality of full color holographic projection system, this
paper analyzes the effect of the fill rate and the color crosstalk on the reconstruction image quality. Transverse and
longitudinal chromatisms are removed by resampling the object information and loading a specially designed virtual
phase distribution in the computer hologram respectively. We proposed time sequence updating chart of RGB laser to
solve the problem of color crosstalk. Experimental results are also provided to verify the improvement of the image
quality.
Hologram, as a type of diffractive optical element, is sensitive to wavelength in the process of optoelectronic reconstruction. Due to the different wavelengths of three prime colors used in full color holographic projections, there are chromatisms which badly spoil the reconstructed image. The chromatisms are composed of transverse and longitudinal chromatisms. For a computer generated hologram, transverse chromatism can be compensated by resampling the object information. However, it becomes more complex for longitudinal chromatism. This paper analyzes how the image is reconstructed from the phase-type hologram in a holographic projection system and the causations of longitudinal chromatism. To remove it, this paper proposes loading a specially designed phase distribution on a phase-type hologram. The advantage of this method is that it can be achieved by computer calculation and without adding any hardware such as achromatic optical element. A time-sharing system for a full-color hologram projection is developed in this paper. Comparisons have been made between the reconstructed images with and without chromatism. The experimental result shows that the method is effective in removing longitudinal chromatism and the quality of the reconstructed image is improved.
To form a laser image, there usually are two kinds of image forming methods. One is to direct slender laser beams to
special points by light directing element (LDE). The other one is light blocking method where amplitude spatial light
modulator (SLM) used to form laser image. The former method has high light efficiency for none light is blocked,
however precision mechanical instrument used as LDE is needed. In this paper, we describe a new kind of laser imaging
method based on computational holography. Coherent light is used as source and phase SLM as LDE. The function of
phase SLM is to change the phase of light illuminated on it. The light on the pixel will travel in new direction according
to the phase change. Iterative Fourier Transform Algorithm (IFTA), also known as G-S algorithm, is used to find phase
change value of each pixel. The process of laser image forming is simulated with VirtualLab™ and optical system is
built to form the laser image. 2D and 3D image can be formed by the proposed method and they enjoy the advantages of
laser display, such as full color gamut and sharp intensity.
In order to avoid missing low-frequency information of three-dimensional (3D) object, pseudorandom phase is generally
added to object planes in tomographic hologram calculation, but speckle noise is inevitably introduced. A novel method
is proposed to improving the quality of image reconstructed from kinoform generated by computational tomographic
holography. Multi-kinoforms of 3D object are calculated, and the images of these kinoforms are superposed to suppress
the speckle noise of reconstructed image. An invariant pseudorandom phase is added to each object plane in calculating
each kinoform, and the pseudorandom phase is generated randomly again in next kinoform calculating. Digital
reconstruction results from single kinoform and multi-kinoforms are analyzed by considering the peak signal-to-noise
ratio (PSNR) and correlation coefficient (CC) between reconstructed images and original object, which verifies that the
speckle noise in reconstructed image can be suppressed by superposing enough images reconstructed from the kinoforms.
Electro-optical reconstruction results also confirm that the method is available.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.