模态输入方式对AIGC辅助汽车造型设计的创造力影响研究
STUDY ON THE INFLUENCE OF MODAL INPUT METHODS ON CREATIVITY IN AIGC-ASSISTED AUTOMOTIVE STYLING DESIGN
李卓 1潘跃1
作者信息
摘要
目的:生成式人工智能的技术突破为传统汽车造型设计带来了新的发展机遇.探索AIGC最有效的模态输入方式,将有助于提高人工智能在汽车造型设计中的辅助效果.方法:本研究采用三种模态输入方式:文字、图片、文字+图片进行设计实验.招募36名具备汽车造型设计经验的专业参与者对AIGC生成的内容进行评价,使用了包括创意性、表达力、实用性、惊喜度、美学价值以及品牌融合等六个因素的量表来评估生成内容的创造力水平.结果:实验结果表明,在AIGC辅助汽车造型设计中,模态输入方式显著影响生成内容的创造力水平,其中文字+图片的模态输入方式明显优于文字或图片单独使用的方式,表现出更高的创造力水平.结论:这一发现强调了多模态输入的重要性,有助于激发人工智能模型生成创造性的设计内容.研究对人工智能辅助汽车造型设计实践具有参考价值.
Abstract
Purpose:Breakthroughs in generative artificial intelligence technology present new development opportunities for traditional automotive styling design.Exploring the most effective modal input method for AIGC(AI Generated Content)can help to enhance its auxiliary effect in automotive styling design.Methods:This study employed three modal input methods:text,image,and text+image for design experiments.Thirty-six professional participants with experience in automotive styling design were recruited to evaluate the content generated by AIGC using a scale comprising six factors:creativity,expressiveness,practicality,surprise factor,aesthetic value,and brand integration,to assess the level of creativity in the generated content.Results:The experimental results indicate that modal input methods significantly influence the level of creativity in the generated content when AIGC assists in automotive styling design.Among these,the text+image modal input method significantly outperforms using text or image alone,exhibiting a higher level of creativity.Conclusion:This finding underscores the importance of multimodal input,aiding in eliciting creatively designed content from artificial intelligence models.The study holds reference value for the practice of artificial intelligence-assisted automotive styling design.
关键词
AIGC/人智交互/模态输入方式/设计辅助/汽车造型Key words
AIGC/Human-AI interaction/modal input methods/design assistance/automotive styling引用本文复制引用
基金项目
湖北省重点研发计划项目(2022BAA071)
湖北省高等学校省级教学研究项目(2022122)
出版年
2024
NETL
NSTL
万方数据