Abstract:The ability to present three-dimensional (3D) scenes with continuous depth sensation has a profound impact on virtual and augmented reality, human–computer interaction, education and training. Computer-generated holography (CGH) enables high-spatio-angular-resolution 3D projection via numerical simulation of diffraction and interference<a href="/articles/s41586-020-03152-0#ref-CR1">1</a>. Yet, existing physically based methods fail to produce holograms with both per-pixel focal control and accurate occlusion<a href="/articles/s41586-020-03152-0#ref-CR2">2</a>,<a href="/articles/s41586-020-03152-0#ref-CR3">3</a>. The computationally taxing Fresnel diffraction simulation further places an explicit trade-off between image quality and runtime, making dynamic holography impractical<a href="/articles/s41586-020-03152-0#ref-CR4">4</a>. Here we demonstrate a deep-learning-based CGH pipeline capable of synthesizing a photorealistic colour 3D hologram from a single RGB-depth image in real time. Our convolutional neural network (CNN) is extremely memory efficient (below 620 kilobytes) and runs at 60 hertz for a resolution of 1,920 × 1,080 pixels on a single consumer-grade graphics processing unit. Leveraging low-power on-device artificial intelligence acceleration chips, our CNN also runs interactively on mobile (iPhone 11 Pro at 1.1 hertz) and edge (Google Edge TPU at 2.0 hertz) devices, promising real-time performance in future-generation virtual and augmented-reality mobile headsets. We enable this pipeline by introducing a large-scale CGH dataset (MIT-CGH-4K) with 4,000 pairs of RGB-depth images and corresponding 3D holograms. Our CNN is trained with differentiable wave-based loss functions<a href="/articles/s41586-020-03152-0#ref-CR5">5</a> and physically approximates Fresnel diffraction. With an anti-aliasing phase-only encoding method, we experimentally demonstrate speckle-free, natural-looking, high-resolution 3D holograms. Our learning-based approach and the Fresnel hologram dataset will help to unlock the full potential of holography and enable applications in metasurface design<a href="/articles/s41586-020-03152-0#ref-CR6">6</a>,<a href="/articles/s41586-020-03152-0#ref-CR7">7</a>, optical and acoustic tweezer-based microscopic manipulation<a href="#ref-CR8">8</a>,<a href="#ref-CR9">9</a>,<a href="/articles/s41586-020-03152-0#ref-CR10">10</a>, holographic microscopy<a href="/articles/s41586-020-03152-0#ref-CR11">11</a> and single-exposure volumetric 3D printing<a href="/articles/s41586-020-03152-0#ref-CR12">12</a>,<a href="/articles/s41586-020-03152-0#ref-CR13">13</a>.

Calculating real-time phase-only holograms through autoencoder neural network

High-speed Computer-Generated Holography Using an Autoencoder-Based Deep Neural Network

High-fidelity, Model-Driven Deep Learning Network for Phase-Only Computer-Generated Holography (conference Presentation)

Real-time 4K computer-generated hologram based on encoding conventional neural network with learned layered phase

Unsupervised Fourier-inspired neural network for real-time and high-fidelity computer-generated holography

Phase Dual-Resolution Networks for a Computer-Generated Hologram

Real-time High-Quality Computer-Generated Hologram Using Complex-Valued Convolutional Neural Network

Phase-only Hologram Generated by a Convolutional Neural Network Trained Using Low-Frequency Mixed Noise.

Generating high-quality phase-only holograms of binary images using global loss and stochastic homogenization training strategy

54.2: High‐fidelity model‐driven deep learning network for phase‐only computer‐generated holography

Fourier-inspired Neural Module for Real-Time and High-Fidelity Computer-Generated Holography.

Efficient Computer-Generated Holography Based on Mixed Linear Convolutional Neural Networks

Conditional neural holography: a distance-adaptive CGH generator

Vision Transformer-Based, High-Fidelity, Computer-Generated Holography

End-to-end learning of 3D phase-only holograms for holographic display

Deep-learning-aided multi-focal hologram generation

Towards real-time photorealistic 3D holography with deep neural networks

Progress of the Computer-Generated Holography Based on Deep Learning

Computer Holography Using Deep Neural Network with Fourier Basis.

Non-iterative Phase Hologram Generation with Optimized Phase Modulation.

High Bandwidth-Utilization Digital Holographic Reconstruction Using an Untrained Neural Network