图书情报工作

图书情报工作 >

2026 , Vol. 70 >Issue 5: 56 - 68

DOI: https://doi.org/10.13266/j.issn.0252-3116.2026.05.005

研究论文

基于AIGC的文化遗产图像智能生成方法研究——以青铜器为例

  • 侯西龙 ,
  • 王雨鑫 ,
  • 臧君涵 ,
  • 刘文斌 ,
  • 王晓光
展开
  • 1 曲阜师范大学传媒学院 日照 276826;
    2 武汉大学文化遗产智能计算实验室 武汉 430072;
    3 武汉大学信息管理学院 武汉 430072
侯西龙，副教授，博士，硕士生导师；王雨鑫，硕士研究生；臧君涵，硕士研究生；刘文斌，博士研究生；王晓光，教授，博士生导师，通信作者，E-mail: wxguang@whu.edu.cn。

收稿日期: 2025-01-06

  修回日期: 2025-02-18

  网络出版日期: 2026-03-04

基金资助

本文系国家社会科学基金青年项目“文化遗产图像资源关联数据集成与叙事化呈现研究”（项目编号：21CTQ041）研究成果之一。

收起

Research on Intelligent Generation Methods for Cultural Heritage Images Based on AIGC: A Case Study of Chinese Bronze Ware

  • Hou Xilong ,
  • Wang Yuxin ,
  • Zang Junhan ,
  • Liu Wenbin ,
  • Wang Xiaoguang
Expand
  • 1 School of Communication, Qufu Normal University, Rizhao 276826;
    2 Intelligent Computing Laboratory for Cultural Heritage, Wuhan University, Wuhan 430072;
    3 School of Information Management, Wuhan University, Wuhan 430072
Hou Xilong, Associate Professor, PhD, Master’s Supervisor; WangYuxin, Master’s Candidate, Zang Junhan, Master’s Candidate; Liu Wenbin, Doctoral Candidate; Wang Xiaoguang, Professor, PhD, Doctoral Supervisor, Corresponding Author, E-mail: wxguang@whu.edu.cn.

Received date: 2025-01-06

  Revised date: 2025-02-18

  Online published: 2026-03-04

Supported by

This work is supported by the youth project of National Social Science Fund of China, titled “Linked Data Integration and Narrative Presentation of Cultural Heritage Image Resources” (Grant No.21CTQ041).

Fold

摘要

[目的/意义] 在AIGC赋能文化遗产领域知识生产范式变革的背景下，研究旨在探索生成式AI在文化遗产领域图像智能生成及其创新性应用，以促进数智时代文化遗产资源的活化利用与创新发展。[方法/过程] 以青铜器风格图像生成为例，首先设计构建青铜器知识本体模型，规范描述多维度的青铜器知识体系；其次，提出大语言模型增强的人智协同图像语义标注方法，利用LoRA算法对图像生成模型进行训练和微调，实现青铜器风格图像的智能生成与创作。最后，总结提出AIGC驱动的文化遗产领域图像智能计算和活化利用的逻辑与路径。[结果/结论] 通过对稳定扩散模型的微调，增强青铜器图像生成的质量及图文一致性，为青铜器文物保护与文化传播提供新的工具与技术支持。AIGC技术为文化遗产领域图像智能理解与计算提供了强大支持，助力领域智慧数据资源建设，并推动图像资源的创意转化与活化利用。

关键词： 图像生成; AIGC; 文化遗产; LoRA; 知识组织; 青铜器

本文引用格式

侯西龙 , 王雨鑫 , 臧君涵 , 刘文斌 , 王晓光 . 基于AIGC的文化遗产图像智能生成方法研究——以青铜器为例[J]. 图书情报工作, 2026 , 70(5) : 56 -68 . DOI: 10.13266/j.issn.0252-3116.2026.05.005

Abstract

[Purpose/Significance] In the context of AIGC, the paradigm of knowledge production within cultural heritage domain has been transformed. This study aims to explore the application of generative AI in intelligent image generation and innovative practice for cultural heritage to promote the revitalization and preservation of cultural heritage in the digital intelligence era. [Method/Process] This paper focused on the image generation of Chinese bronze ware style. Firstly, it designed and constructed an ontology for Chinese bronze ware, standardized its description of multi-dimensional knowledge. Secondly, it proposed a human-AI collaboration image semantic annotation method enhanced by large language models. Utilizing the LoRA (Low-Rank Adaptation) algorithm, it trained and fine-tuned image generation model to achieve intelligent creation of Chinese bronze ware style images. Finally, it summarized and outlined the logic and pathways for AIGC-driven intelligent computation and revitalization applications of cultural heritage image resources. [Result/Conclusion] By fine-tuning the stable diffusion model, the quality of Chinese bronze ware image generation and the consistency between images and text has been enhanced. This provides new tools and technical support for the preservation and cultural dissemination of Chinese bronze ware. AIGC technology offers strong support for intelligence understanding and computation of cultural heritage images, facilitates the construction of smart data resources, and promotes the creative transformation and revitalization of cultural heritage images resources.

Key words： image generation; AIGC; cultural heritage; low-rank adaptation; knowledge organization; Chinese bronze ware

参考文献

[1] CHEN H, XIANG Q, HU J, et al. Comprehensive exploration of diffusion models in image generation: a survey[J]. Artificial intelligence review, 2025, 58(4): 99.
[2] MA Z Y, ZHANG Y Z, JIA G L, et al. Efficient diffusion models: A comprehensive survey from principles to practices[J]. arXiv preprint, 2024, arXiv:2410.11795.
[3] LI Z, ZHANG J, LIN Q, et al. Hunyuan-DiT: A powerful multi-resolution diffusion transformer with fine-grained Chinese understanding[J]. arXiv preprint, 2024, arXiv:2405.08748.
[4] 王晓光. AI时代文化遗产数据资源建设及其开放应用[J]. 文献与数据学报, 2024(2): 15-16. (WANG X G. Construction and open application of cultural heritage data resources in AI era[J]. Journal of library and data, 2024(2): 15-16.)
[5] 童祁. 生成式AI时代, 人工智能如何影响文博保护与传承[EB/OL]. [2025-02-10] https://mp.weixin.qq.com/s/4Z0v76a54_CBXcnjGILnkA. (TONG Q. In the era of generative AI, how does AI affect the protection and inheritance of cultural relics [EB/OL]. [2025-02-10] https://mp.weixin.qq.com/s/4Z0v76a54_CBXcnjGILnkA.)
[6] WU S, CHEN Z, SU Q. Knowledge-aware artifact image synthesis with LLM-enhanced prompting and multi-source supervision[C]// Proceedings of the 32nd ACM international conference on multimedia. New York: Association for computing machinery, 2024: 2719-2728.
[7] GU Y, FANG X, DENG X, et al. A new Chinese landscape paintings generation model based on Stable Diffusion using DreamBooth[J]. arXiv preprint, 2024, arXiv:2408.08561.
[8] 侯云鹏, 彭涵, 刘育晖. 基于LoRA模型的非遗数字化传承:以楚漆器为例[J]. 设计艺术研究, 2024, 14(1): 14-18. (HOU Y P, PENG H, LIU Y H. Digital inheritance of intangible cultural heritage based on the LoRA model: a case study of Chu lacquerware[J]. Design research, 2024, 14(1): 14-18.)
[9] 于鹏, 张毅. 从特征辨识到图像生成：基于AIGC范式的苗族服饰设计[J]. 丝绸, 2024, 61(3): 1-10. (YU P, ZHANG Y. From feature recognition to image generation: Miao ethnic costume design based on the AIGC paradigm[J]. Journal of silk, 2024, 61(3): 1-10.)
[10] 李莉, 毛子晗, 吕思奇, 等. GAN与Diffusion在传统纹样设计中的实验研究[J]. 丝绸, 2024, 61(8): 9-22. (LI L, MAO Z H, LV S Q, et al. An experimental study on the application of GAN and diffusion models in traditional pattern design[J]. Journal of silk, 2024, 61(8): 9-22.)
[11] 纪向宏, 夏宇滋. “非遗+AI”模式下的湖南土特产包装创新设计[J]. 包装工程, 2024, 45(16): 301-312. (JI X H, XIA Y Z. Innovation design of Hunan local product packaging under the model of “intangible cultural heritage + AI”[J]. Packing engineering, 2024, 45(16): 301-312.)
[12] 高永杰, 吕欣. 生成式AI技术进化与图像艺术生产范式革新[J]. 现代传播(中国传媒大学学报), 2023, 45(9): 159-168. (GAO Y J, LV X. Evolution of generative AI technology and the innovation of image art production paradigm[J]. Modern communication (Journal of communication university of China), 2023, 45(9): 159-168.)
[13] 李春桃, 戚睿华, 杨溪, 等. 基于深度学习技术的青铜鼎分期断代研究[J]. 出土文献, 2023(3): 16-32, 154-155. (LI C T, QI R H, YANG X, et al. Periodization of bronze ding-vessels based on deep learning[J]. Excavated documents, 2023(3): 16-32, 154-155.)
[14] WANG, W, LU Z. Few-shot bronze vessel classification via siamese fourier networks[J]. Scientific reports, 2024, 14(1): 18011.
[15] ZHANG X. AI?assisted restoration of yangshao painted pottery using LoRA and stable diffusion[J]. Heritage, 2024, 7: 6282-6309.
[16] GUAN H S, YANG H X, WANG X Y, et al. Deciphering oracle bone language with diffusion models[J]. arXiv preprint, 2024, arXiv:2406.00684.
[17] WANG L, LOU Y, LI X, et al. GlyphGenius: Unleashing the potential of AIGC in Chinese character learning[J]. IEEE access, 2024, 12: 136420-136434.
[18] 台北故宫博物院. 故宫典藏资源检索[EB/OL]. [2025-02-03]. https://digitalarchive.npm.gov.tw/Integrate/Index. (Taipei palace museum. Taipei palace museum collection [EB/OL]. [2025-02-03]. https://digitalarchive.npm.gov.tw/Integrate/Index.)
[19] WEI T, CHEN Y. A ding ontology of Chinese bronze[J]. Journal on computing and cultural heritage, 2023, 1: 1-12.
[20] 胡汗林, 邓三鸿. 知识图谱在青铜器数字馆藏建设中的应用[J]. 数字图书馆论坛, 2023, 19(4): 1-8. (HU H L, DENG S H. Application of knowledge graph in the construction of bronze digital collection[J]. Digital library forum, 2023, 19(4): 1-8.)
[21] GILARDI F, ALIZADEH M, KUBLI M. Chatgpt outperforms crowd-workers for text-annotation tasks[J]. arXiv preprint, 2023, arXiv:2303.15056.
[22] HU E J, SHEN Y, WALLIS P, et al. LoRA: Low-rank adaptation of large language models[J]. arXiv preprint, 2021, arXiv:2106.09685.
[23] KOHYA S. sd-scripts [EB/OL]. [2025-02-18]. https://github.com/kohya-ss/sd-scripts.
[24] HARTWIG S, ENGEL D, SICK L, et al. A survey on quality metrics for text-to-image generation[J]. arXiv preprint, 2024, arXiv:2403.11821.
[25] KIRSTAIN Y, POLYAK A, SINGER U, et.al. Pick-a-Pic: an open dataset of user preferences for text-to-image generation[J]. arXiv preprint, 2023, arXiv:2305.01569.
[26] XU J LIU X, WU Y, et al. Imagereward: Learning and evaluating human preferences for text-to-image generation[J]. arXiv preprint, 2023, arXiv:2304.05977.
[27] HESSEL J, HOLTZMAN A, FORBES M, et al. Clipscore: A reference-free evaluation metric for image captioning[J]. arXiv preprint, 2021, arXiv:2104.08718.
[28] 马乐存, 詹希旎, 朱齐宇, 等.AIGC驱动的GLAM数智融合创新发展研究[J]. 农业图书情报学报, 2023, 35(5): 4-15. (MA L C, ZAN X N, ZHU Q Y, et al. Digital intelligence integration innovation development of GLAM driven by AIGC[J]. Journal of library and information science in agriculture, 2023, 35(5): 4-15.)
[29] 王遂川, 庞欢, 熊亦冉, 等. 图像数智化:现象及其阐释[J]. 浙江社会科学, 2023(3): 85-93, 84, 159. (WANG S H, PANG H, XIONG Y R, et al. Digital smart images: phenomenon and interpretation[J]. Zhejiang social science, 2023(3): 85-93, 84, 159.)
[30] 范炜, 曾蕾. AI新时代面向文化遗产活化利用的智慧数据生成路径研究[J]. 中国图书馆学报, 2024, 50(2): 4-29. (FAN W, ZENG L. Exploring the generation path of smart data for the activation and utilization of cultural heritage in the new ear of AI[J]. Journal of library science in China, 2024, 50(2): 4-29.)
[31] 陈涛, 刘炜, 孙逊, 等. IIIF与AI作用下的文化遗产应用研究新模态[J]. 中国图书馆学报, 2021, 47(2): 67-78. (CHEN T, LIU W, SUN X, et al. New modality of applied research on cultural heritage under the action of IIIF and AI[J]. Journal of library science in China, 2021, 47(2): 67-78.)
[32] 敦煌研究院. 数字敦煌开放素材库[EB/OL]. [2025-02-01]. https://ip.e-dunhuang.com/. (Dunhuang academy. Digital dunhuang open material library [EB/OL]. [2025-02-01]. https://ip.e-dunhuang.com/.)
[33] 腾讯研究院. 中国文化遗产数字化研究报告2023-2024[EB/OL]. [2025-02-10]. https://mp.weixin.qq.com/s/rf3Oq3oEa2Xd4oAAN8cDew. (Tencent research institute. Research report on the digitization of Chinese cultural heritage 2023-2024[EB/OL]. [2025-02-10]. https://mp.weixin.qq.com/s/rf3Oq3oEa2Xd4oAAN8cDew.)
[34] 王晓光, 侯西龙. 面向活化利用的文化遗产智慧数据建设论纲[J]. 信息资源管理学报, 2023, 13(5): 4-14, 43. (WANG X G, HOU X L. The construction of cultural heritage smart data for the inheritance and activation[J]. Journal of information resources management, 2023, 13(5): 4-14, 43.)
[35] ZHANG W, ZHANG JW, WONG KK, et al. Computational approaches for traditional Chinese painting: From the “six principles of painting” perspective[J]. Journal of computer science and technology, 2024, 39(2): 269-285.
[36] JAILLANT L. Introduction to the special issue: using visual AI applied to digital archives[J]. Digital humanities quarterly, 2024, 18(2).
[37] 贾爱玲, 张玮, 张建伟, 等. 识印鉴章:中国书画印章分析系统[J]. 计算机辅助设计与图形学学报, 2025, 37(9): 1526-1536. (JIA A L, ZHANG W, ZHANG J W, et al. Identifying and authenticating seals: Chinese calligraphy and painting seal analysis system[J]. Journal of computer aided design & computer graphics, 2025, 37(9): 1526-1536.)
[38] 林馨怡, 吴泓嘉, 袁稚庭, 等.基于图像提取与修复的计算机辅助古画印章分析[J]. 计算机辅助设计与图形学学报, 2025, 37(2): 254-264. (LIN X Y, WU H J, YUAN Z T, et al. Computer aided analysis of ancient painting seals based on image extraction and restoration[J]. Journal of computer aided design & computer graphics, 2025, 37(2): 254-264.)
[39] BIN Y, SHI W, DING Y, et al. GalleryGPT: Analyzing paintings with large multimodal models[C]// Proceedings of the 32nd ACM international conference on multimedia, New York: Association for computing machinery, 2024: 7734-7743.
[40] ZHANG W, KAM-KWAI W, XU B, et al. CultiVerse: Towards cross-cultural understanding for paintings with large language model[J]. arXiv preprint, 2024, arXiv:2405.00435.
[41] XU Z, ZHANG X, CHEN W, et al. MuralDiff: Diffusion for ancient murals restoration on large-scale pre-training[J]. IEEE transactions on emerging topics in computational intelligence, 2024, 8(3): 2169-2181.
[42] SMITS T, WEVERS M. Multimodal turn in digital humanities. Using contrastive machine learning models to explore, enrich, and analyze digital visual historical collections[J]. Digital scholarship in the humanities, 2023, 38:1267-1280.
[43] 韩茹雪, 饶梓欣, 许鑫. AIGC赋能的GLAM机构内容生产研究[J]. 图书情报工作, 2024, 68(22): 4-15. (HAN R X, RAO Z X, XU X. A study of content production in GLAM empowered by AIGC[J]. Library and information service, 2024, 68(22): 4-15.)
[44] WANG B, CHEN Q, WANG Z. Diffusion-based visual art creation: A survey and new perspectives[J]. arXiv preprint, 2024, arXiv:2408.12128.
[45] ZHAO S, HUANG Y F, He X Y, et al. Reviving mural art through generative AI: A comparative study of AI-generated and hand-crafted recreations[C]//CHI conference on human factors in computing systems (CHI ’25). New York: ACM, 2025.
[46] 清华大学美术学院.《韩熙载夜宴图》AI沉浸式体验展[EB/OL]. [2025-01-26]. https://mp.weixin.qq.com/s/igYdA2JsPbbgA_FOkeSBLw. (Academy of art & design, Tsinghua university. AI immersive experience exhibition of Han Xizai’s Night Banquet [EB/OL]. [2025-01-26]. https://mp.weixin.qq.com/s/igYdA2JsPbbgA_FOkeSBLw.)
[47] Harvard Art Museum. About the AI explorer [EB/OL]. [2025-01-25]. https://ai.harvardartmuseums.org/about.
[48] Cleveland Museum of Art. AI: ArtLens for slack [EB/OL]. [2025-01-25] https://www.clevelandart.org/digital-innovations/artlens-for-slack.
Options
文章导航

/

〈 〉