Controllable Generation of Single Blue Calico Patterns Using Stable Diffusion Model

Artificial intelligence (AI) technologies are increasingly permeating various industries, and technological advancements have also brought new opportunities to traditional craftsmanship. This study focuses on the application of artificial intelligence-generated content (AIGC) to assist traditional printing and dyeing techniques, aiming to address the generation of new patterns for blue calico by constructing an end-to-end low-loss pattern generation framework. Through analyzing the stylistic characteristics of blue calico patterns and comparing how the Low-Rank Adaptation (LoRA) model optimizes the base model of Stable Diffusion, as well as the structural control effects of ControlNet on images, this paper ultimately establishes a single-pattern generation pathway for blue calico. This pathway integrates multi-model control based on LoRA and ControlNet with the generation mechanism of the Stable Diffusion model. Validation results demonstrate that patterns generated using this approach expand design content while preserving the intrinsic features of blue calico, thereby providing a digitally innovative solution that balances cultural heritage and technical innovation for the preservation of intangible cultural heritage in traditional printing and dyeing craftsmanship.