Research Progress of Panoramic Dynamic Holographic Three-Dimensional Display Technology (Invited)

Significance Humans live in a world surrounded by three-dimensional (3D) information, perceiving 3D scenes from different viewing angles through motion. Panoramic 3D display technology can offer 360° viewing angle, with natural, comprehensive, and immersive naked-eye 3D visual experience, which is the key to achieving true 3D visual interaction. Panoramic dynamic holographic display technology not only achieves a full horizontal 360° viewing angle and smooth motion parallax but also provides all depth cues required by human vision, achieving the closest reconstruction effect to real 3D objects. As a result, it has become a shared focus of both academia and industry. Furthermore, panoramic dynamic holographic 3D display technology demonstrates broad application prospects in fields such as entertainment display, surgical navigation, battlefield simulation, and holographic communication. It has emerged as a key technological pathway driving the development of naked-eye holographic 3D displays toward practical implementation. Progress The fundamental framework of panoramic dynamic holographic display technology is summarized, consisting of four core steps: 360° RGB-depth (RGB-D) data acquisition, viewing-angle computer-generated hologram (CGH) computation, viewing-angle CGHs optimization, and panoramic dynamic holographic display system construction. This framework constitutes a comprehensive optimization problem that spans the entire data pipeline, from acquisition and processing to transmission and final display. In the aspect of 360° 3D information acquisition, software rendering techniques for virtual scenes have become relatively mature. For real scenes, with the support of consumer-grade depth cameras, researchers have developed various flexible solutions such as single depth camera acquisition and multi depth camera array synchronous acquisition methods to adapt to different scene types. Regarding viewing-angle CGHs computation, a variety of primitive-based algorithms (e.g., point-based, line-based, plane-based, and layer-based methods) have been proposed based on diffraction physics models. Furthermore, to improve efficiency, research has focused on technical approaches including look-up table (LUT), wavefront recording plane (WRP), motion compensation, and deep learning. With the aid of parallel computing devices such as graphics processing unit (GPU), real-time generation of a limited number of viewing-angle CGHs has been preliminarily achieved. In terms of viewing-angle CGHs optimization, methods such as iterative algorithms, error diffusion, and single-sideband filtering have been employed to effectively suppress noise introduced during CGH generation and encoding, thereby improving reconstruction quality. Moreover, deep learning-based CGH compression techniques have surpassed the performance limitations of traditional compression schemes, offering the potential for efficient transmission of massive holographic data. Regarding display systems, time-multiplexing schemes based on digital micromirror devices (DMDs) have demonstrated clear advantages, enabling preliminary realization of panoramic dynamic 3D display with a 360° horizontal viewing angle, high viewing angle resolution, and high frame rate. Conclusions and Prospects Although considerable progress has been made, the road to practical implementation of panoramic dynamic holographic 3D display technology is hindered by several critical challenges. These encompass the end-to-end optimization of complex real scenes from data acquisition through display, the real-time computational demands of generating massive viewing-angle CGHs, and the ongoing need for system integration and miniaturization. In the future, with coordinated advances in algorithms, hardware, and system design, panoramic holographic display technology is expected to integrate deeply with artificial intelligence, thereby establishing a novel content generation and interaction paradigm of “what you think is what you see”. Through multimodal understanding and real-time rendering, abstract inputs such as natural language or hand-drawn sketches could be directly transformed into high-quality spatial 3D image, enabling revolutionary immersive interactive experiences in education, entertainment, design, and remote collaboration.