EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

3D digital reconstruction and presentation of traditional textile arts of the li people using deep learning technology

Chunnan Cao, Lei Zhang, Xiang Qian

Discover Applied Sciences · 22.05.2026 · 8 (8)

DOI: 10.1007/s42452-026-08869-7 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Discover Applied Sciences, online erschienen am 22.05.2026. Verfasst von Chunnan Cao und 2 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Hainan Provincial Natural Science Foundation of China “Research on the Digital Revitalization Presentation of Hainan Li Brocade Traditional Craftsmanship”.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The traditional textile arts of the Li people, known for their complex weaving, dyeing, and embroidery patterns, represent a vital component of China’s intangible cultural heritage. However, the fragile condition of many artifacts and the lack of digital documentation hinder their preservation and public dissemination. This research proposes a Deep Learning (DL) driven 3D digital reconstruction and presentation framework to preserve and visualize the Li people’s textile heritage in immersive environments. The Li heritage textile 3D reconstruction dataset with 30 Li textile images is obtained. The raw images underwent preprocessing using bilateral filtering. A Convolutional Neural Network (CNN)-based feature extractor and a 3D reconstruction network are employed to recreate surface geometry and fine details of textile structures. The proposed model, Snap-Drift Cuckoo Search-driven Three-dimensional Generative Adversarial Networks (SDCS-3D-GAN) model, aims to generate effective 3D textures while maintaining the original color and weave continuity. The 3D-GAN provides an efficient textile structure with consistent color and weaving, while the SDCS optimization is used to fine-tune the hyperparameters in 3D-GAN for more significant performance. Experimental results demonstrate that the SDCS-3D-GAN achieved superior texture effectiveness and visual fidelity compared to existing DL models with accuracy (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:0.982$$\end{document}), precision (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:0.984$$\end{document}), recall (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:0.989$$\end{document}), F1-score (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:0.951$$\end{document}), Structural Similarity Index Measure (SSIM-\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:0.95$$\end{document}), and Peak-Signal-to-Noise Ratio (PSNR-\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:35.89$$\end{document} dB) by employing a Python implementation. Finally, the reconstructed textiles were integrated into a virtual presentation platform, enabling immersive AR/VR-based visualization, evaluated through user interaction preparations. This research provides a sustainable digital preservation pathway that merges Li cultural artistry with modern computational innovation.

Förderung

  • Hainan Provincial Natural Science Foundation of China “Research on the Digital Revitalization Presentation of Hainan Li Brocade Traditional Craftsmanship”

Zitieren

APA

Cao, C., Zhang, L., & Qian, X. (2026). 3D digital reconstruction and presentation of traditional textile arts of the li people using deep learning technology. Discover Applied Sciences, 8(8). https://doi.org/10.1007/s42452-026-08869-7

Vancouver

Cao C, Zhang L, Qian X. 3D digital reconstruction and presentation of traditional textile arts of the li people using deep learning technology. Discover Applied Sciences. 2026;8(8). doi:10.1007/s42452-026-08869-7

BibTeX
@article{Cao2026,
  author = {Cao, Chunnan and Zhang, Lei and Qian, Xiang},
  title = {3D digital reconstruction and presentation of traditional textile arts of the li people using deep learning technology},
  journal = {Discover Applied Sciences},
  year = {2026},
  volume = {8},
  number = {8},
  doi = {10.1007/s42452-026-08869-7}
}
Kostenlos starten →