Aug 2026· AI in Civil Engineering· Vol 5· 0 citations· 78 references
TL;DR
This study proposes an AI-driven generative design workflow that translates semantic inputs into 2D imagery and 3D models, enabling the systematic learning and replication of stylistic features from a quintessential southern Chinese architectural ornament—the Lingnan stucco relief.
Abstract
The accelerating digitalization of the construction sector has brought into sharp focus the inefficiencies and labor-intensive nature inherent in traditional craftsmanship for historical architecture. In response, this study proposes an AI-driven generative design workflow that translates semantic inputs into 2D imagery and 3D models, enabling the systematic learning and replication of stylistic features from a quintessential southern Chinese architectural ornament—the Lingnan stucco relief. This approach moves beyond the conventional reliance on individual expertise and heuristic empirical methods. The structural performance of the AI-generated designs was evaluated through combined material testing and numerical simulations. Concurrently, hybrid additive manufacturing techniques were explored for rapid physical prototyping. Collectively, this integrated framework illustrates the feasibility of revitalizing traditional craft practices within a contemporary intelligent paradigm, offering a technological pathway for the preservation and renovation of historic built heritage.
Immersive visualization can support interpretation of architectural heritage in historical paintings, yet translating 2D pictorial evidence into navigable 3D scenes remains challenging. Conventional workflows rely on physical survey data, while direct generative AI (GenAI) may produce structural hallucinations and lack historical constraints. This study proposes a human-in-the-loop GenAI-assisted framework for producing immersive 3D visualization prototypes rather than historically verified reconstructions. It integrates multi-view image generation, knowledge-informed review, single-image-to-3D generation, topology inspection, and perceptual calibration. Four fragments from the Northern Song Dynasty painting Along the River During the Qingming Festival were examined as a single-case proof of concept. Across three tested model pairs, raw AI assets were generated in approximately 3–4 min and were suitable for distant-background use; close-up visualization required 1–2 h of refinement, while basic structural editability required 4–5 h of post-processing, reducing the initial time advantage. A mixed-methods study with nine domain experts and 30 non-expert participants used the UES-SF, an adapted VisAWI, and semi-structured interviews analyzed through inductive thematic analysis. All eight subscale scores exceeded their neutral midpoints after Bonferroni correction (all adjusted p<0.001), indicating favorable perceptions of the guided experience. Interviews suggested potential for spatial exploration, museum interpretation, and education. However, geometric discontinuities, detail loss, color deviation, and historical-semantic errors remained, requiring expert review and manual correction. Transferability beyond this artwork and architectural tradition remains untested.
As AIGC (AI-generated content) technology develops at high speed, AIGC participates in design from different dimensions. In 3D design, generative modeling largely relieves the pressure of high-intensity repetitive work; however, the precision of generative models is not comparable to 3D models routinely designed by designers. Based on this situation, this study takes the digital design practice of Minnan patterned tile as an example to compare the similarities and differences between the traditional 3D modeling workflow and the AIGC-assisted modeling workflow. This paper holds that differences in output forms, final quality, and entry barriers between AIGC modeling and traditional modeling mean that the two 3D design approaches are not in an oppositional relationship; we can take the essence of each and combine their advantages to carry out 3D design work, and a human-led, AI-assisted collaborative mode is bound to be an important trend in the future innovative development of 3D digital design.
As a representative of cultural heritage, Suzhou Ming-style furniture embodies a composite knowledge system integrating material properties, structure, and craftsmanship. While Generative AI shows potential in creative design, current models lack an intrinsic understanding of physical and manufacturing constraints, often producing unproducible visual illusions. To bridge the gap between pixelated images and actual production, this paper proposes and develops a production-aware AI co-creation system that translates implicit traditional craftsmanship into explicit digital rules. Utilizing a node-based workflow and a knowledge-embedded estimation module, the system connects creative exploration with production configuration. It generates multimodal outputs, including renderings, structural exploded views, and practical cost and timeline estimations. Preliminary evaluations indicate high consistency with actual workshop production experience, effectively lowering the creative threshold for non-professionals. This research demonstrates how AI transcends mere visual generation to become a new pathway for the dynamic preservation and revitalization of cultural heritage.
Liwen Fan, Aojie Feng, Yuxuan Li et al.· Creativity & Cognition· 0 citations
In the context of the digital economy era, the digital preservation and innovation of traditional apparel crafts face dual challenges in cultural dissemination and technological integration. This study proposed and validates an integrated digital workflow that incorporates generative AI and physical simulation integrating artificial intelligence and high-fidelity physical simulation to facilitate the contemporary transformation of traditional apparel culture. A technical framework of “intelligent pattern generation — parametric structural modeling — dynamic physical simulation — virtual-real integration” was constructed, with the costumes of the nationally recognized intangible cultural heritage, Nuo Opera, serving as an empirical case. By building a dedicated dataset and employing lightweight fine-tuning techniques, intelligent derivation and stylized control of Nuo mask patterns were achieved, alongside three-dimensional simulation of the structural characteristics and dynamic aesthetics of traditional costumes. Experimental results indicate that the patterns generated by this method exhibit significantly higher visual complexity compared to conventional design approaches, while the design process is effectively streamlined. The physical simulation demonstrated commendable performance in both garment form stability and color fidelity of cultural elements. This study demonstrated that the workflow can support cross-media dissemination and innovative application of ICH craftsmanship through digital twin systems, providing a reference solution with both theoretical and practical value for the intelligent upgrading of the apparel industry and cultural heritage preservation.
Huafeng Feng, Jiajia Li, C. Keyi· Journal on Computing and Cul...· 0 citations
This work presents an alternative, result-oriented, data-driven method based on generative AI to assist engineers in the conceptual design phase of bridge construction, and demonstrates that result-oriented, data-driven generative models can support early-stage bridge topology exploration under controlled conditions.
Aswin Lal, Chongjie Kang, B. Özcan et al.· Journal of computing in civi...· 0 citations
We present a lightweight production pipeline that transforms fragile historical costumes into dynamic digital assets using Generative Re-imagination to bridge the gap between restricted physical archives and production requirements. Our key insight is that labor-intensivke physical scanning and multi-camera photogrammetric capture can be replaced by AI-generated virtual views, which serve as the foundation for 3DGS model creation. We introduce a novel method to unify these partial 3DGS models into seamless temporal representations (4DGS), termed Blended 3DGS, and demonstrate the framework's versatility in the context of digital actor pipelines for costume design and facial animation, as well as virtual period fashion for Virtual Productions.
B. Takács, Patrik Pencz, Zsuzsanna Vincze· SIGGRAPH Posters· 0 citations