关键词: 产品感性设计, 感性工学, 风格语义, 产品风格设计

Abstract: To address the issues of unexplainability and unpredictability in existing intelligent SUV style generation research,an intelligent framework integrating visual cognition,Kansei engineering and AI technology was proposed for SUV style generation.Following the extraction of visual perception data from an SUV image dataset,quantification theory type I was introduced to establish an explainable mapping relationship between styling features and style imagery,thereby constructing an explainable style prediction model.Utilizing StyleGAN3 to generate a diverse stylistic spectrum of SUV models,and a generative framework was developed under the dual constraints of an interpretable style prediction model and the detection of key SUV styling elements,resulting in an interpretable solution that incorporates predefined style prediction scores.The results demonstrated that the generated images achieved an FID50k_full score of 7.12,a styling element detection accuracy of 92%,and an 82.5% compliance rate within the style prediction error threshold (±0.5).This intelligent framework provided interpretability and predictability for SUV automotive style generation,effectively reducing designers'workload in the conceptual design stage and enhancing the practical value of design outcomes.

Key words: Kansei product design, Kansei engineering, style semantics, product style design