Jun 08, 2023
The state
Light Publishing Center, Changchun Institute of Optics, Fine Mechanics And Physics, CAS image: a, Diagram of a point-cloud CGH algorithm, showing typical axis conventions. Every point of the ship
Light Publishing Center, Changchun Institute of Optics, Fine Mechanics And Physics, CAS
image: a, Diagram of a point-cloud CGH algorithm, showing typical axis conventions. Every point of the ship model point cloud creates a point-spread function on the hologram plane (color code: hue corresponds to phase, brightness to amplitude). The superposition of all point-spread functions yields the final hologram. b, Diagram of layer-based CGH, showing how a 3D object is segmented into layers, where every scene element is assigned to its closest intermediate wavefront. Here, an exemplary point with its PSF influence region on the matching layer is shown. All layers are accumulated into a hologram using e.g., convolutional numerical propagation. view more
Credit: by David Blinder, Tobias Birnbaum, Tomoyoshi Ito, Tomoyoshi Shimobaba
Holography is a methodology based on coherent light that can fully describe the optical field. It is a two-step method, consisting of an interferometric recording and a reconstruction step. Its ability to capture, measure and reproduce any given wavefield, has made it useful in a broad range of applications. Examples are digital holographic microscopy, surface characterization of complex objects, particle image velocimetry, and the visualization of 3D content. An important challenge in this context is computer generated holography (CGH), i.e., the modeling of numerical diffraction, calculating how light propagates through space and interacts with materials. CGH is highly computationally intensive, requiring specialized algorithms and hardware for the accurate and efficient calculation of holograms.
In a new paper published in Light Science & Application, resulting from a collaboration between researchers from the Department of Electronics and Informatics at the Vrije Universiteit Brussel and IMEC (Belgium), and the Graduate School of Engineering, Chiba University (Japan), they present a broad overview of multiple aspects of the state-of-the-art in CGH. They present a classification of modern CGH algorithms, algorithmic CGH acceleration techniques, the latest dedicated hardware solutions and perceptual quality evaluation.
CGH algorithms are classified and compared based on the elements that represent them. Point cloud methods discretize the objects into a finite collection of discrete luminous points. 3D objects can also be decomposed into a relatively small number of geometric primitives & basis functions whose diffraction pattern can be efficiently computed. Polygon methods encode triangles as wavefield pieces, leveraging the fact that diffraction between planes can be efficiently computed using convolutions. Layer based methods slice the 3D scene into depth layers, assigning scene elements to their closest layer, causing the relative proximity to a virtual plane to limit diffraction spreading and thereby enhancing spatial locality. Ray-tracing methods approximate the hologram by a discretized light field allowing for the use of conventional computer graphics software for rendering, followed by a conversion of the rays to small wavefront segments.
In addition, most CGH algorithms can be sped up by means of acceleration techniques. Examples are the use of sparsity, whereby a signal can be modelled by a relatively small number of significant coefficients when expressed in the right transform space, e.g., using wavefront recording planes, holographic stereograms and coefficient shrinking. Other examples involve the use of look-up tables, dynamic CGH acceleration for holographic videos and deep-learning based acceleration.
Furthermore, software-hardware co-design necessary to optimize CGH calculation. The review covers optimizations and caching in CPUs and GPUs, low precision and low bit-width calculations, and an analysis of multiple Field-programmable gate arrays and application-specific integrated circuit systems for CGH.
Another important consideration are the capabilities and limitations of current holographic display technologies, and how to compensate for them. This includes comparing the properties of spatial light modulators, forms of complex-amplitude encoding, image restoration and speckle reduction algorithms.
Finally, the authors discuss visual quality assessment in order to evaluate the quality of generated holograms, so to allow for the optimization of perceptual quality. This involves both subjective quality assessment, where various distorted and ground-truth reconstructions are presented in a controlled environment to many human observers assigning scores, and objective quality assessment, entailing the use of mathematical functions to quantify the accuracy of numerically reconstructed wavefields.
Tremendous advances in CGH algorithms in terms of both visual quality and computational complexity have been made, especially in recent years. Despite these optimizations, the computation of high-resolution, wide viewing angle photo-realistic digital holograms in real-time remains an important challenge to be tackled. Nevertheless, existing real-time CGH implementations show that driving high quality holographic displays is already feasible and ahead of current displaying hardware capabilities in several respects.
The authors believe that a close integration of CGH algorithms with dedicated hardware systems will likely yield the biggest computational gains, enabling immersive holographic television.
Light: Advanced Manufacturing
10.37188/lam.2022.035
Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.
image: a, Diagram of a point-cloud CGH algorithm, showing typical axis conventions. Every point of the ship model point cloud creates a point-spread function on the hologram plane (color code: hue corresponds to phase, brightness to amplitude). The superposition of all point-spread functions yields the final hologram. b, Diagram of layer-based CGH, showing how a 3D object is segmented into layers, where every scene element is assigned to its closest intermediate wavefront. Here, an exemplary point with its PSF influence region on the matching layer is shown. All layers are accumulated into a hologram using e.g., convolutional numerical propagation.computer generated holography (CGH)Point cloud methodsgeometric primitives & basis functions Polygon methods Layer based methods Ray-tracingsparsitylook-up tablesdynamic CGH accelerationdeep-learningbased acceleration.software-hardwareco-designcaching low precisionlow bit-widthField-programmable gate arraysapplication-specific integrated circuit comparing the properties of spatial light modulators,complex-amplitude encodingrestorationspeckle reductionvisual quality assessmentsubjective quality assessment objective quality assessmentDisclaimer: