文献类型：期刊文献

英文题名：Deep brand analogical design system: Enhancing creativity and brand style consistency for LoRA-based automotive style transfer

作者：Chen, Yumiao[1];Yan, Yifan[1];Ruan, Huanhuan[1];Yu, Chunyang[2]

机构：[1]East China Univ Sci & Technol, Sch Art Design & Media, Shanghai 200237, Peoples R China;[2]China Acad Art, Design AI Lab, Hangzhou, Peoples R China

年份：2026

卷号：86

起止页码：158

外文期刊名：JOURNAL OF MANUFACTURING SYSTEMS

收录：;EI(收录号：20261120285922);WOS:【SCI-EXPANDED(收录号:WOS:001726790300001)】；

基金：This study was partly supported by the Humanities and Social Sci-ence Fund of Ministry of Education of the People's Republic of China (No.24YJA760013) , the National Natural Science Foundation of China (No.51905175, No.52475248) , and Xie Youbai Design Science Research Foundation (No.XYB-DS-202301) .

语种：英文

外文关键词：Stable Diffusion model; Brand styling aesthetics; Analogical design; Intelligent generation

摘要：Under the paradigm-shifting impact of artificial intelligence (AI) on product design, balancing brand consistency with creative styling has emerged as a critical competitive strategy in automotive design. This study proposes a human-AI hybrid intelligent design methodology integrating analogical reasoning theory and Stable Diffusion models. We develop a framework spanning data processing, model training, design generation, and evaluation. Based on this framework, we establish a system that addresses traditional limitations in efficiency and creativity while advancing automotive design from experience-driven to AI-driven practices. The research begins with the construction of a multi-brand automotive image dataset and the training of brand-specific Stable Diffusion models, with the objective of capturing inherited design elements and establishing data-model foundations. Subsequently, ControlNet-enabled multi-distance analogical design generation achieves semantic mapping of morphological features and details. The research evaluates the brand style consistency of generative design results, demonstrating the potential of analogical design in maintaining brand identity. The contribution of this study lies in the interdisciplinary integration of generative AI and analogical reasoning, and in the proposal of a human-AI collaborative design framework for brand-centric styling exploration. This demonstrates potential in reconciling styling design and brand inheritance within product appearance variation.