目的 传承保护和创新发展土家族非物质文化遗产——西兰卡普传统织锦工艺,助力解决传统手工艺风格复原难、图案创新效率低等问题。方法 提出一种通过LoRA模型微调Stable Diffusion的智能生成方法,通过对西兰卡普图纹形象、构成布局、色彩运用三大特征维度的深度解析,构建精标数据集,并利用LoRA模型对Stable Diffusion进行精准微调,以完成少样本风格迁移实验。结果 通过艺术与技术的结合,不仅实现了西兰卡普风格的精确捕捉与复现,还完成了高质量图案的智能生成。结论 在西兰卡普的特征提取与智能生成方面,通过现代设计方法与智能生成技术的深度融合,使其在顺应现代发展趋势、完成活态传承任务的同时,为同类研究提供了创新性的思路与方法。
Abstract
The work aims to inherit, protect and innovate in the intangible cultural heritage of the Tujia ethnic group - the traditional weaving craft of Xilan Kapu, and help solve the problems of difficult restoration of traditional handicraft styles and low efficiency of pattern innovation. An intelligent generation method for fine-tuning Stable Diffusion through LoRA model was proposed. Through deep analysis of the three major feature dimensions of Xilan Kapu patterns, including image, composition layout, and color application, a precise dataset was constructed, and the LoRA model was used for precise fine-tuning of Stable Diffusion to complete the small sample style transfer experiment. The combination of art and technology not only had the precise capture and reproduction of the Xilan Kapu style, but also completed the intelligent generation of high-quality patterns. In terms of feature extraction and intelligent generation in Xilan Kapu, the deep integration of modern design methods and intelligent generation technology has enabled it to adapt to modern development trends and complete the task of dynamic inheritance, while providing innovative ideas and methods for similar research.
关键词
西兰卡普 /
Stable Diffusion /
LoRA /
少样本风格迁移 /
智能生成
Key words
Xilan Kapu /
Stable Diffusion /
LoRA /
small sample style transfer /
intelligent generation
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 邓丽. 土家族织锦艺术在湖北民族学院美术教学中的实践探索研究[D]. 北京: 中央民族大学, 2010.
DENG L.Research on the Practice of Tujia Brocade Art in Art Teaching of Hubei University for Nationalities[D]. Beijing: Central University for Nationalities, 2010.
[2] 陆垚. 湘西土家织锦艺术传承与保护研究——评《湘西土家族织锦技艺》[J]. 毛纺科技, 2020, 48(5): 111-112.
LU Y.Study on Inheritance and Protection of Tujia Brocade Art in Xiangxi—Comment on Tujia Brocade Skills in Xiangxi[J]. Wool Textile Journal, 2020, 48(5): 111-112.
[3] 乔锦浩, 肖懿, 崔誉丹, 等. 特定风格剪纸艺术的智能生成[J]. 包装工程, 2021, 42(14): 74-80.
QIAO J H, XIAO Y, CUI Y D, et al.Intelligent Generation for Specific Style of Paper-Cutting Art[J]. Packaging Engineering, 2021, 42(14): 74-80.
[4] 蔡妮娜. 鄂西土家族传统图形研究及启示[D]. 武汉: 武汉理工大学, 2007.
CAI N N.Research on Tujia Traditional Graphics in Western Hubei and Its Enlightenment[D]. Wuhan: Wuhan University of Technology, 2007.
[5] 王敏, 彭昌盛. 符号学视阈下土家族西兰卡普文化因子提取及设计应用研究[J]. 包装工程, 2024, 45(20): 333-345.
WANG M, PENG C S.Extraction and Design Application of Tujia Xilankapu Cultural Factors from the Perspective of Semiotics[J]. Packaging Engineering, 2024, 45(20): 333-345.
[6] 贺景卫. 湘西土家织锦矢量图典[M]. 长沙: 湖南师范大学出版社, 2017.
HE J W.Vector Atlas of Tujia Brocade in Xiangxi[M]. Changsha: Hunan Normal University Press, 2017.
[7] 苏晓. 浅析湘西土家织锦的图案构成特征[J]. 艺术科技, 2016, 29(4): 75.
SU X.Analysis on the Pattern Composition Characteristics of Tujia Brocade in Xiangxi[J]. Art Science and Technology, 2016, 29(4): 75.
[8] 程莎, 叶洪光. 土家织锦图纹浅析[J]. 武汉纺织大学学报, 2013, 26(4): 21-23.
CHENG S, YE H G.Analysis of Tujia Brocade Pattern[J]. Journal of Wuhan Textile University, 2013, 26(4): 21-23.
[9] 麻丹龄. 土家织锦纹样骨格罗列及创新应用研究[D]. 武汉: 中南民族大学, 2020.
MA D L.A Study on the Listing and Innovative Application of Tujia Brocade Patterns[D]. Wuhan: South- central University for Nationalities, 2020.
[10] 赖小静. 西兰卡普图案特征及其成因探究[J]. 现代装饰(理论), 2017(1): 214-215.
LAI X J.Study on the Characteristics and Causes of Xilankapu Pattern[J]. Modern Decoration (Theory), 2017(1): 214-215.
[11] 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.
[12] CROITORU F A, HONDRU V, IONESCU R T, et al.Diffusion Models in Vision: A Survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(9): 10850-10869.
[13] HO J, JAIN A, ABBEEL P.Denoising Diffusion Probabilistic Models[J]. Advances in Neural Information Processing Systems, 2020, 33: 6840-6851.
[14] HARTLEY Z K J, LIND R J, POUND M P, et al. Domain Targeted Synthetic Plant Style Transfer Using Stable Diffusion, LoRA and ControlNet[C]//2024 IEEE/ CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). Seattle: IEEE, 2024.
[15] BORHANIFARD Z, FAILI H, YAGHOOBZADEH Y.Combining Replay and LoRA for Continual Learning in Natural Language Understanding[J]. Computer Speech & Language, 2025, 90: 101737.
基金
国家社科基金艺术学一般项目（21BG120）
PDF(7439 KB)
渝公网安备 50010702501717号