Wenbin Zhou

Wenbin Zhou

香港大学，博士研究生

The University of Hong Kong, Ph.D. Candidate

我目前是香港大学（HKU）电机电子工程系（EEE）三年级博士生，现于由彭祎帆（Evan Peng）教授领导的计算成像与混合表征实验室（WeLight@HKU）开展研究工作。

我于中国科学技术大学（USTC）获得物理学学士学位，并辅修计算机科学，师从刘利刚教授。随后，我在香港大学攻读并获得计算机科学硕士学位。此外，我曾在普渡大学担任研究助理，在 Bedrich Benes 教授指导下开展研究，并作为访问学者在加州大学伯克利分校（UC Berkeley）开展科研工作，由 Brian Barsky 教授接待与指导。

我的研究兴趣包括计算全息、光学、计算机图形学，以及虚拟现实与增强现实。

I am a third-year Ph.D. student in the Department of Electrical and Electronic Engineering (EEE) at The University of Hong Kong (HKU). I am currently conducting research in the Computational Imaging & Mixed Representation Laboratory (WeLight@HKU), led by Prof. Yifan (Evan) Peng.

I received my bachelor’s degree in Physics, with a minor in Computer Science, from the University of Science and Technology of China (USTC), under the supervision of Prof. Ligang Liu. I then pursued a master’s degree in Computer Science at HKU. In addition, I gained research experience as a Research Assistant at Purdue University, working with Prof. Bedrich Benes, and as a visiting researcher at UC Berkeley, hosted by Prof. Brian Barsky.

My research interests include computational holography, optics, computer graphics, and virtual/augmented reality.

题目

深度学习辅助的计算机生成全息（CGH)

摘要：

全息技术正日益受到关注，被认为是虚拟现实（VR）与增强现实（AR）下一代显示系统的重要技术基础。实现照片级真实感的三维视觉内容，对于提升感知真实感与用户沉浸体验至关重要。然而，在这些应用场景中，现有计算机生成全息（CGH）方法仍受到多方面限制，包括计算效率不足、对数据吞吐量要求高，以及对关键深度线索复现不完整等问题。本文介绍了近年来深度学习赋能的全息显示技术进展，重点讨论其如何缓解上述核心瓶颈。

Title

Deep Learning-aided Computer-Generated Holography

Abstract：

Holographic techniques have gained growing attention as a promising foundation for next-generation display systems in virtual reality (VR) and augmented reality (AR). Achieving photorealistic three-dimensional visual content is essential for enhancing perceptual realism and user immersion. Nevertheless, prevailing computer-generated holography (CGH) approaches in these contexts remain limited by insufficient computational efficiency, high data throughput requirements, and incomplete reproduction of key depth cues. This manuscript reviews recent advances in deep learning–enabled holographic display technologies aimed at addressing these fundamental constraints.