TY - JOUR
T1 - Compact freeform near-eye display system design enabled by optical-digital joint optimization
AU - Xu, Huiming
AU - Yang, Tong
AU - Cheng, Dewen
AU - Wang, Yongtian
N1 - Publisher Copyright:
Copyright © 2024 Xu, Yang, Cheng and Wang.
PY - 2024
Y1 - 2024
N2 - The near-eye display (NED) systems, designed to project content into the human eye, are pivotal in the realms of augmented reality (AR) and virtual reality (VR), offering users immersive experiences. A small volume is the key for a fashionable, easy-to-wear, comfortable NED system for industrial and consumer use. Freeform surfaces can significantly reduce the system volume and weight while improving the system specifications. However, great challenges still exist in further reducing the volume of near-eye display systems as there is also a limit when using only freeform optics. This paper introduces a novel method for designing compact freeform NED systems through a powerful optical–digital joint design. The method integrates a geometrical freeform optical design with deep learning of an image compensation neural network, addressing off-axis nonsymmetric structures with complex freeform surfaces. A design example is presented to demonstrate the effectiveness of the proposed method. Specifically, the volume of a freeform NED system is reduced by approximately 63% compared to the system designed by the traditional method, while still maintaining high-quality display performance. The proposed method opens a new pathway for the design of a next-generation ultra-compact NED system.
AB - The near-eye display (NED) systems, designed to project content into the human eye, are pivotal in the realms of augmented reality (AR) and virtual reality (VR), offering users immersive experiences. A small volume is the key for a fashionable, easy-to-wear, comfortable NED system for industrial and consumer use. Freeform surfaces can significantly reduce the system volume and weight while improving the system specifications. However, great challenges still exist in further reducing the volume of near-eye display systems as there is also a limit when using only freeform optics. This paper introduces a novel method for designing compact freeform NED systems through a powerful optical–digital joint design. The method integrates a geometrical freeform optical design with deep learning of an image compensation neural network, addressing off-axis nonsymmetric structures with complex freeform surfaces. A design example is presented to demonstrate the effectiveness of the proposed method. Specifically, the volume of a freeform NED system is reduced by approximately 63% compared to the system designed by the traditional method, while still maintaining high-quality display performance. The proposed method opens a new pathway for the design of a next-generation ultra-compact NED system.
KW - compact structure
KW - freeform optics
KW - image compensation network
KW - near-eye display system
KW - optical design
KW - optical–digital joint optimization
UR - http://www.scopus.com/inward/record.url?scp=85202812397&partnerID=8YFLogxK
U2 - 10.3389/fphy.2024.1440129
DO - 10.3389/fphy.2024.1440129
M3 - Article
AN - SCOPUS:85202812397
SN - 2296-424X
VL - 12
JO - Frontiers in Physics
JF - Frontiers in Physics
M1 - 1440129
ER -