Using the Stable Diffusion Model to Achieve Ink Painting Style Transfer and Contemporary Art Re-Creation
Y. Liu, L. Liu, Z. Q. Liu · Advanced Electromagnetics · 2026
Chinese ink painting is regarded as a jewel of Eastern art, with its aesthetic of “living” and the “integration of form and spirit” manifested through techniques of hooking, kneading, and leaving blank. Existing artificial intelligence (AI) methods often reproduce visual forms while failing to preserve intrinsic artistic expression. To address this limitation, this paper proposes a hierarchical style transfer framework based on Stable Diffusion that aligns abstract aesthetic concepts such as the balance between emptiness and reality through CLIP, while LoRA fine-tuning and ControlNet preserve structural consistency and enhance expressive flexibility under ink-edge constraints. Considering the growing importance of intelligent visual information processing in electromagnetic sensing and multimodal communication systems, the proposed framework also provides a reference for interpretable semantic representation and digital cultural information transmission. Experimental results demonstrate improved structural fidelity (SSIM 0.78) and wellcontrolled blank-space distribution (28.5%). The framework not only facilitates digital artistic creation but also contributes to the preservation and revitalization of traditional crafts, including weaving and dyeing, through explainable AI technologies with potential value for cross-domain intelligent information processing.