计算全息作为一种三维显示手段，能够基于衍射计算实现对目标光场的精确重建，在元宇宙通讯、AR/VR头戴显示、车载抬头显示等方向均有着重要的应用。如何实现高速且高质量的相位全息图生成是计算全息领域发展的关键问题，也是当前该方向的重要研究课题。近年来，深度学习技术的飞跃式发展为上述问题的解决提供了一条新的技术路径。本文介绍了计算全息技术的基本原理及算法分类，综述了近年来所提出的基于深度学习的计算全息解决方案，比较了各类方案的优势与不足，展望了深度学习技术在计算全息领域的发展与挑战。

波前编码过程将计算全息所得的复振幅波前变换为与显示器件匹配的调制函数，是计算全息显示的关键技术之一。现有的计算全息显示器件大多只能实现单一振幅或单一相位调制，因此需要将复振幅波前编码为相应的振幅型或相位型全息图。本文围绕基于液晶空间光调制器的计算全息显示，综述了相位优化编码与复振幅转化编码的基本原理与算法步骤，分析了常见的波前编码方案框架，针对不同编码方法的适用范围进行讨论，为计算全息图波前编码提供方法选择参考。

相比二维显示，三维显示可以提供更接近真实世界的图像内容，是5G、大数据、元宇宙和物联网领域的关键性技术。计算全息三维显示可以提供所有种类的深度线索，被认为是三维显示的终极实现方式，在智能制造、远程教育、异地办公和娱乐社交等领域都具有广阔的应用前景。文中首先介绍了计算全息技术的发展历史和重要技术节点，分析了高质量全息动态三维显示面临的挑战，主要包括计算全息图重建质量不足、波前调制器件和全息显示系统性能受限以及三维内容源匮乏等，进一步介绍了已有的解决方案，比较了各类方案的优势与不足，进而分析了高质量全息动态三维显示的主要研究方向，包括低噪声全息图的获取、像质优化和畸变校正技术的开发以及三维内容源的构建。实现高质量、低噪声、无畸变、高刷新、真三维的动态全息三维显示，是计算全息显示发展的必由之路，也是元宇宙等典型应用对全息三维显示提出的必然要求。

The bidirectional error diffusion (BERD) algorithm is free from random phase modulation that introduces speckle noise on the reconstructed images, compared with other computer-generated phase-only hologram (POH) approaches. During the POH generation process, the amplitudes of all pixels are traditionally set to one for diffusing the errors to their neighborhood of unprocessed pixels. In this paper, we reveal that the reconstruction quality depends on the uniform amplitude value for different object pattern. The pattern-adaptive BERD (PA-BERD) algorithm is proposed for high-quality holographic reconstruction. The optimized amplitude value can be acquired for each object pattern and each propagation distance. The PA-BERD-based POHs have shown higher reconstruction quality than traditional BERD-based POHs in simulations as well as optical experiments.

This Letter describes an approach to encode complex-amplitude light waves with spatiotemporal double-phase holograms (DPHs) for overcoming the limit of the space-bandwidth product (SBP) delivered by existing methods. To construct DPHs, two spatially macro-pixel encoded phase components are employed in the SBP-preserved resampling of complex holograms. Four generated sub-DPHs are displayed sequentially in time for high-quality holographic image reconstruction without reducing the image size or discarding any image terms when the DPHs are interweaved. The reconstructed holographic images contain more details and less speckle noise, with their signal-to-noise ratio and structure similarity index being improved by 14.64% and 78.79%, respectively.

In this review, the principle and the optical methods for light-field display are introduced. The light-field display is divided into three categories, including the layer-based method, projector-based method, and integral imaging method. The principle, characteristic, history, and advanced research results of each method are also reviewed. The advantages of light-field display are discussed by comparing it with other display technologies including binocular stereoscopic display, volumetric three-dimensional display, and holographic display.