WGCNet: wavelet global context network for speckle-free 4K holographic display with high-frequency preservation
Photonics Research · 01.07.2026 · 14 (7) · 2925
DOI: 10.1364/prj.588876 ↗Kurz gefasst
Fachartikel in der Zeitschrift Photonics Research, online erschienen am 22.06.2026. Verfasst von Yingjun Luo und 3 weiteren Autor:innen. Die Arbeit nennt 60 Literaturangaben. Gefördert von: National Natural Science Foundation of China.
Abstract (Originalsprache)
Neural-network-based computer-generated holography (CGH) has been extensively confirmed to break the tradeoff between algorithm runtime and reconstruction quality. However, conventional U-Net-based CGH methods lose high-frequency detail due to inadequate global context modeling, which constrains the enhancement of realism and immersion in holographic displays. To address these challenges, we propose WGCNet: a two-stage wavelet global context network for generating speckle-free, high-fidelity 4K phase-only holograms (POHs). This framework integrates a two-dimensional (2D) wavelet down-sampling technique to enhance the U-Net backbone network and leverages physical prior knowledge to preserve high-frequency details. A lightweight global attention module is introduced to model long-range dependencies. We demonstrate that the proposed model achieves a peak signal-to-noise ratio of 38.68 dB and a structural similarity index of 0.9615 on the DIV2K dataset at 4K resolution. The method significantly reduces the speckle noise in reconstruction while effectively preserving fine details. This synergistic approach establishes an effective solution for high-resolution holographic displays, offering potential for enhancing visual realism and immersion in virtual reality (VR) and augmented reality (AR) applications.
Förderung
- National Natural Science Foundation of China
Zitieren
APA
Luo, Y., Liu, K., He, Z., & Cao, L. (2026). WGCNet: wavelet global context network for speckle-free 4K holographic display with high-frequency preservation. Photonics Research, 14(7), 2925. https://doi.org/10.1364/prj.588876
Vancouver
Luo Y, Liu K, He Z, Cao L. WGCNet: wavelet global context network for speckle-free 4K holographic display with high-frequency preservation. Photonics Research. 2026;14(7):2925. doi:10.1364/prj.588876
BibTeX
@article{Luo2026,
author = {Luo, Yingjun and Liu, Kexuan and He, Zehao and Cao, Liangcai},
title = {WGCNet: wavelet global context network for speckle-free 4K holographic display with high-frequency preservation},
journal = {Photonics Research},
year = {2026},
volume = {14},
number = {7},
pages = {2925},
doi = {10.1364/prj.588876}
}
EUVIMED