Design method for reconstruction of new Chinese-style fur clothing based on the FLUX model
Source abstract
Abstract
Purpose To address the insufficient application of traditional Chinese-style fur elements in modern new Chinese-style design and issues of poor detail retention, semantic deviation, and overfitting in existing style transfer methods based on Stable Diffusion, this study proposes a FLUX-based reconstruction method for Chinese-style fur clothing to realize modern reconstruction and high-quality style transfer of traditional elements. Design/methodology/approach With FLUX as the core, integrate FLUX-Low-Rank Adaptation (LoRA) and FLUX-Redux; analyze Chinese-style fur design elements, build a fashion fur dataset, train the FLUX-LoRA model, perform conditional encoding of fur elements via Redux, and verify feasibility through qualitative experiments, Structural Similarity Index, and Peak Signal-to-Noise Ratio. Findings The FLUX-LoRA model generates text-matching fashion fur designs with good generalization; Redux enables natural fusion of traditional elements. This method outperforms IP-Adapter and StableDiffusion_v1.5+LoRA in detail, consistency, and realism. Originality/value Combining FLUX with LoRA and Redux for Chinese-style fur reconstruction solves issues of traditional Stable Diffusion, provides a feasible path for applying Chinese-style fur in modern new Chinese-style design, and contributes to protecting and inheriting Chinese traditional fur clothing culture.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Generative Adversarial Networks and Image Synthesis · Face recognition and analysis · 3D Shape Modeling and Analysis
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Xi'an Polytechnic University , ,
Jiangnan University , ,
Funding
Humanities and Social Sciences Research Project of the Ministry of Education
References (43)
Show source-provided references
1. AI-assisted inheritance of Qinghua porcelain cultural genes and sustainable design using low-rank adaptation and stable diffusion10.3390/electronics14040725 ↗
2. Enhancing the digital inheritance and development of Chinese intangible cultural heritage paper-cutting through stable diffusion LoRA models10.3390/app142311032 ↗
3. Scaling rectified flow transformers for high-resolution image synthesis10.48550/arXiv.2403.03206 ↗
4. A neural algorithm of artistic style10.1167/16.12.326 ↗
5. Generative adversarial nets10.1145/3422622 ↗
6. CCST-GAN: generative adversarial networks for Chinese calligraphy style transfer
7. Denoising diffusion probabilistic models10.48550/arxiv.2006.11239 ↗
8. CLIP-GAN: stacking CLIPs and GAN for efficient and controllable text-to-image synthesis10.1109/TMM.2025.3535304 ↗
9. LoRA: low-rank adaptation of large language models10.48550/arXiv.2106.09685 ↗
10. Adaptive stylized image generation for traditional Miao Batik using style-conditioned LCM-LoRA enhanced diffusion models10.3390/math13121947 ↗
11. Image-to-image translation with conditional adversarial networks10.1155/2020/8894309 ↗
12. Jin, Z. and Song, Z. (2023), “Generating coherent comic with rich story using ChatGPT and stable diffusion”, ArXiv,2305.11067v2, doi: 10.48550/arXiv.2305.11067, available at: https://arxiv.org/abs/2305.11067v2
13. Auto-encoding variational bayes10.48550/arXiv.1312.6114 ↗
14. Kuru, J. (2023), “Training non-typical character models for stable diffusion utilizing open source AIS”, The University of Arizona, available at: https://repository.arizona.edu/handle/10150/668639
15. Clipstyler: image style transfer with a single text condition
16. Labs, B.F., Batifol, S., Blattmann, A., Boesel, F., Consul, S., Diagne, C., Dockhorn, T., English, J., English, Z., Esser, P., Kulal, S., Lacey, K., Levi, Y., Li, C., Lorenz, D., Müller, J., Podell, D., Rombach, R., Saini, H., Sauer, A. and Smith, K. (2025), “FLUX. 1 Kontext: flow matching for in-context image generation and editing in latent space”, arXiv preprint arXiv:2506.15742, doi: 10.48550/arXiv.2506.15742.
17. Generation and design of Helan mountain petroglyphs based on style transfer algorithm
18. Style transfer of Chinese traditional dermatoglyphic pattern with wide diffusion model
19. Efficient generation of digital rock CT images using LoRA-enhanced stable diffusion models10.1016/j.ige.2025.06.002 ↗
20. Flow matching for generative modeling10.48550/arXiv.2210.02747 ↗
21. Podell, D., English, Z., Lacey, K., Blattmann, A., Dockhorn, T., Müller, J., Penna, J. and Rombach, R. (2023), “SDXL: improving latent diffusion models for high-resolution image synthesis”, arXiv preprint arXiv:2307.01952, doi: 10.48550/arXiv.2307.01952.
22. Learning transferable visual models from natural language supervision10.48550/arXiv.2103.00020 ↗
23. Exploring the limits of transfer learning with a unified text-to-text transformer10.48550/arXiv.1910.10683 ↗
24. High-resolution image synthesis with latent diffusion models10.48550/arXiv.2112.10752 ↗
25. U-net: convolutional networks for biomedical image segmentation10.48550/arXiv.1505.04597 ↗
26. Dreambooth: fine tuning text-to-image diffusion models for subject-driven generation10.48550/arXiv.2208.12242 ↗
27. Singh, A.K. and Patras, I. (2024), “Fashionsd-x: multimodal fashion garment synthesis using latent diffusion”, arXiv preprint arXiv:2404.18591, doi: 10.48550/arXiv.2404.18591.
28. Batik clothes auto-fashion using conditional generative adversarial network and U-Net10.1109/EIConCIT50028.2021.9431867 ↗
29. MLP-mixer: an all-MLP architecture for vision10.48550/arXiv.2105.01601 ↗
30. KurtaGen: LoRA-enhanced diffusion for Indian traditional garment design generation and manipulation
31. Attention is all you need10.48550/arXiv.1706.03762 ↗
32. Design of watercolor cultural and creative products based on style transfer algorithm10.1155/2022/2711861 ↗
33. Image localized style transfer to design clothes based on CNN and interactive segmentation10.1155/2020/8894309 ↗
34. Constructing a personalized AI assistant for shear wall layout using stable diffusion10.48550/arXiv.2305.10830 ↗
35. ClothGAN:generation of fashionable dunhuang clothes using generative adversarial networks10.1080/09540091.2020.1822780 ↗
36. Region-aware style transfer between thangka images via combined segmentation and adaptive style fusion
37. Yang, C., Liu, C., Deng, X., Kim, D., Mei, X., Shen, X. and Chen, L.-C. (2024), “1.58-bit FLUX”, arXiv preprint arXiv:2412.18653, doi: 10.48550/arXiv.2412.18653.
38. Ye, H., Zhang, J., Liu, S., Han, X. and Yang, W. (2023), “Ip-adapter: text compatible image prompt adapter for text-to-image diffusion models”, arXiv preprint arXiv:2308.06721, doi: 10.48550/arXiv.2308.06721.
39. Sustainable design on intangible cultural heritage: Miao embroidery pattern generation and application based on diffusion models10.3390/su17177657 ↗
40. Adding conditional control to text-to-image diffusion models10.48550/arXiv.2302.05543 ↗
41. DyArtbank: diverse artistic style transfer via pre-trained stable diffusion and dynamic style prompt artbank10.1016/j.knosys.2025.112959 ↗
42. Unpaired image-to-image translation using cycle-consistent adversarial networks10.48550/arXiv.1703.10593 ↗
43. Research on AI painting generation technology based on the [stable diffusion]10.7236/IJASC.2023.12.2.90 ↗