Digital Restoration of Tang Dynasty Clothing Structure and Fiber Material Simulation Based on 3D Reconstruction

Liuyun Zhang , Bo Xia , Fei Ai

Article
2026 / Volume 9 / Pages 1293-1318
Received 27 August 2025; Accepted 4 November 2025; Published 29 April 2026
https://doi.org/10.31881/TLR.2026.1293

Abstract
To address the challenge that individual historical sources, such as a single mural image, provide limited information for high-precision 3D structural restoration and fiber material physical property simulation, this paper proposes a comprehensive method called “Multi-source Fusion and Anisotropic Reconstruction (MFAR)” that integrates multiple heterogeneous data sources to overcome the inherent limitations of any single perspective. The Scale-Invariant Feature Transform–Fast Library for Approximate Nearest Neighbors–Random Sample Consensus (SIFT-FLANN-RANSAC) feature matching strategy integrates multi-source data from murals, pottery figurines, and images of unearthed cultural relics, enhancing the integrity and reliability of the input data. Subsequently, 12 key structural points are detected based on the High-Resolution Network version 2 (HRNet-v2) network to provide semantic support for subsequent geometric modeling. Next, the Poisson reconstruction and As-Rigid-As-Possible (ARAP) deformation algorithm are combined to optimize the 3D mesh topology and achieve high-fidelity geometric restoration of clothing shapes. Additionally, this paper innovatively constructs an 8-chain anisotropic hyperelastic material model to simulate the typical mechanical response and fiber orientation properties of Tang Dynasty silk. Experimental results show that this method reduces the geometric error of a round-necked robe to 3.4 mm, achieves an average root mean square error (RMSE) of 2.7 pixels for keypoint localization, achieves a maximum Fourier spectrum similarity of 0.964, and retains an average historical feature rate of 91.9%. These findings provide a verifiable technical path and methodological support for the digital preservation and cultural heritage of ancient clothing.

Keywords
cultural heritage digitalization, 3D reconstruction, structural restoration, fiber material simulation, Tang Dynasty clothing