Deep learning methods for designing AR display using VHOE-based waveguide
Shiuan‐Huei Lin, Zih-Fan Chen, Chi Sun, Yuan-Yan Liang, Tsung-Hsun Yang, Yeh‐Wei Yu, Ching‐Cherng Sun, Vera Marinova · 2025
We propose a CNN model to simulate wave propagation in an AR optical system using Volume Holographic Optical Elements (VHOEs) as combiners. The architecture and CWT-based training dataset are discussed, focusing on chromatic dispersion-induced color distortion. After training, the model achieves nearly 100% color accuracy with unknown spectra, demonstrating deep learning’s potential for modeling complex optical systems and advancing HOE-based AR displays.