高晨

中国光电信息福建省科技创新实验室，研发人员

Fujian Science & Technology Innovation Laboratory for Optoelectronic Information of China，

Research & Development

高晨，2018年获天津大学学士学位，2023年获浙江大学光学工程博士学位。目前在闽都创新实验室（中国福建光电信息科学与技术创新实验室）新型显示科技创新中心从事AR/VR近眼显示和三维显示研发工作。作为骨干参与多项国家级研发项目和实验室自主部署项目。发表研究论文10余篇，授权发明专利1项。曾获ICDT最佳论文奖和优秀学生论文奖。担任Optics Express、SIGGRAPH、IEEE ISMAR、IEEE VR等期刊和会议审稿人。

Chen Gao, obtained his bachelor’s degree from Tianjin University in 2018 and PhD degree in Optical Engineering from Zhejiang University in 2023. Currently, he is working on AR/VR near-eye display and 3D display R&D at the Advanced Display Sci-Tech Innovation Center of Mindu Innovation Laboratory (Fujian Science & Technology Innovation Laboratory for Optoelectronic Information of China). As a core member, he has participated in several national R&D projects and independently deployed laboratory projects. He has published over 10 research papers and holds 1 invention patent. He has won the Best Paper Award and the Distinguished Student Paper Award at ICDT. He serves as a reviewer for journals and conferences, including Optics Express, SIGGRAPH, IEEE ISMAR, and IEEE VR.

题目

高性能压缩光场3D显示研究

摘要：

压缩光场3D显示具有空间分辨率高、显示深度大的优点，但是可视角小、计算量大。针对以上问题，本次报告介绍了研究团队在压缩光场3D显示的可视角提升和计算量降低方面的工作。将注视点渲染技术引入压缩光场近眼显示中，在保证注视点处聚焦线索质量的前提下大大降低了计算量；利用深度学习生成压缩光场显示的编码图像，在避免了复杂迭代计算的同时改善了显示质量和均匀性；提出了针对加法型压缩光场追踪显示的权重优化算法，通过有效节约显示带宽从而呈现多人同时观看的3D显示内容。最后，基于闪烁融合、双目视差和运动视差等相互独立的人眼视觉特性、充分利用显示面板的高刷新率、高空间分辨率和图像内容相关性，发展结合多指向背光、集成成像、压缩光场的混合裸眼3D显示技术，从而继续提升3D显示的可视角度、显示深度和图像质量。

Title

High Performance Compressive Light Field 3D Displays

Abstract：

Compressive light field 3D display has the advantages of high spatial resolution and large display depth, but it is limited by narrow viewing angle and heavy computation. To address these issues, this presentation introduces our research team's work on improving the viewing angle and reducing the computation of compressive light field 3D display. By introducing foveated rendering into compressive light field near-eye displays, the computation is significantly reduced while providing high-quality focus cues within the fovea. By using deep learning to synthesize encoded images for compressive light field displays, the display quality and uniformity are improved without the need for cumbersome iterations. A weight optimization algorithm for additive light field tracking displays is proposed, which effectively saves display bandwidth to present 3D display content for multiple viewers. Finally, based on the independent visual characteristics of human eyes such as flicker fusion, binocular disparity, and motion parallax, and by fully utilizing the high refresh rate, high spatial resolution, and image content correlation of the display panel, a hybrid naked-eye 3D display combining multi-directional backlighting, integral imaging, and compressive light field is developed to further enhance the viewing angle, display depth, and image quality of 3D display.