Product design and evaluation methods based on AI-generated content

doi:10.11996/JG.j.2095-302X.2024061277

Abstract

Abstract:

Generative artificial intelligence (GAI) has become a transformative force in product design, significantly enhancing design efficiency. However, systematic application methods and cases of collaborative multi-type GAI application examples remain scarce. To highlight the innovative role of GAI in product design, a method based on AI-generated content (AIGC) for product form design and evaluation was proposed. First, ChatGPT was utilized to capture the emotional needs of target product users and summarize them into design target imageries. Additionally, ChatGPT served as a prompt generator for Midjourney, generating necessary prompt phrases for the target product. Midjourney constructed a reference library for product forms using these target imageries and prompt phrases. Perceptual questionnaires were then utilized to select distinctive designs as alternatives. Next, the grey relational analysis (GRA) and analytic hierarchy process (AHP) methods were employed to evaluate these alternatives and select the optimal design, with Rhino used to optimize human-machine interaction. Finally, stable diffusion was utilized to quickly generate rendering effects for the optimal design. A case study on electric motorcycles and household vacuum cleaners validated the proposed method. It was found that the collaborative model of multi-type generative AI excelled in analyzing user needs, transforming design concepts, and optimizing design details. This approach revolutionized traditional design processes and improved design efficiency. The proposed method provided product designers with an AIGC-based design approach and established a quantitative evaluation method for AIGC.

Key words: form design, generative AI, design evaluation, grey relationship analysis, analytic hierarchy process

CLC Number:

TB472

LU Peng, WU Fan, TANG Jian. Product design and evaluation methods based on AI-generated content[J]. Journal of Graphics, 2024, 45(6): 1277-1288.

Figures/Tables 23

References 34

[1]	程永胜, 徐骁琪. 用户感性需求驱动下的产品造型意象设计研究[J]. 机械设计与制造, 2023(8): 247-252.
	CHENG Y S, XU X Q. Research on product modeling image design driven by users perceptual demands[J]. Machinery Design & Manufacture, 2023(8): 247-252. (in Chinese)
[2]	张书涛, 张凡, 王永照, 等. 基于形状文法与灰关联的施工升降机造型设计[J]. 机械设计, 2023, 40(7): 157-163.
	ZHANG S T, ZHANG F, WANG Y Z, et al. Modeling design of construction hoist based on shape grammar and grey correlation[J]. Journal of Machine Design, 2023, 40(7): 157-163. (in Chinese)
[3]	MCCORMACK J P, CAGAN J, VOGEL C M. Speaking the Buick language: capturing, understanding, and exploring brand identity with shape grammars[J]. Design Studies, 2004, 25(1): 1-29.
[4]	李永锋, 朱丽萍. 基于结合分析的产品意象造型设计研究[J]. 图学学报, 2012, 33(4): 121-128.
	LI Y F, ZHU L P. Research on form design of product image based on conjoint analysis[J]. Journal of Graphics, 2012, 33(4): 121-128. (in Chinese)
[5]	LU P, HSIAO S W, WU F. A product shape design and evaluation model based on morphology preference and macroscopic shape information[J]. Entropy, 2021, 23(6): 639.
[6]	王剑, 陈丰, 乔印虎. 目标意象约束的人机辅助产品造型进化设计研究[J]. 机械设计, 2022, 39(2): 120-128.
	WANG J, CHEN F, QIAO Y H. Research on evolutionary design of human-machine aided product modeling constrained by target image[J]. Journal of Machine Design, 2022, 39(2): 120-128. (in Chinese)
[7]	HSIAO S W, CHIU F Y, LU S H. Product-form design model based on genetic algorithms[J]. International Journal of Industrial Ergonomics, 2010, 40(3): 237-246.
[8]	林哲辉, 吴正仲, 罗峰, 等. 基于感性工学与人工神经网络的电动剃须刀多感官意象设计方法[J]. 机械设计, 2023, 40(2): 149-156.
	LIN Z H, WU Z Z, LUO F, et al. Multi-sensory design method of electric shavers based on Kansei engineering and artificial neural networks[J]. Journal of Machine Design, 2023, 40(2): 149-156. (in Chinese)
[9]	WU Y, MA L S, YUAN X F, et al. Human-machine hybrid intelligence for the generation of car frontal forms[J]. Advanced Engineering Informatics, 2023, 55: 101906.
[10]	周爱民, 苏建宁, 阎树田, 等. 基于形态美度的产品多意象预测模型[J]. 图学学报, 2018, 39(4): 654-660. DOI
	ZHOU A M, SU J N, YAN S T, et al. Product multi-image prediction model based on aesthetic measure of form[J]. Journal of Graphics, 2018, 39(4): 654-660. (in Chinese) DOI
[11]	LU P, HSIAO S W. A product design method for form and color matching based on aesthetic theory[J]. Advanced Engineering Informatics, 2022, 53: 101702.
[12]	罗仕鉴, 潘云鹤. 产品设计中的感性意象理论、技术与应用研究进展[J]. 机械工程学报, 2007, 43(3): 8-13.
	LUO S J, PAN Y H. Review of theory, key technologies and its application of perceptual image in product design[J]. Chinese Journal of Mechanical Engineering, 2007, 43(3): 8-13. (in Chinese)
[13]	李白杨, 白云, 詹希旎, 等. 人工智能生成内容(AIGC)的技术特征与形态演进[J]. 图书情报知识, 2023, 40(1): 66-74.
	LI B Y, BAI Y, ZHAN X N, et al. The technical features and aromorphosis of artificial intelligence generated content (AIGC)[J]. Documentation, Information & Knowledge, 2023, 40(1): 66-74. (in Chinese)
[14]	WU F, HSIAO S W, LU P. An AIGC-empowered methodology to product color matching design[J]. Displays, 2024, 81: 102623.
[15]	李彦宏. 10年后全世界50%的工作会是提示词工程[J]. 中国商人, 2023(11): 68-69.
	LI Y H. Half of the world's jobs will be prompt engineering in ten years[J]. Chinese Businessman, 2023(11): 68-69. (in Chinese)
[16]	赵文强, 臧欣慈. 皮尔斯符号学视域下AIGC对文创产品的设计探索[J]. 包装工程, 2024, 45(10): 116-126.
	ZHAO W Q, ZANG X C. Design of cultural and creative products based on the combination of pierce semiotics and AIGC[J]. Packaging Engineering, 2024, 45(10): 116-126. (in Chinese)
[17]	李源枫, 王丽, 唐彩云, 等. 基于AIGC辅助技术的智慧交通公共设施设计研究[J]. 包装工程, 2024, 45(6): 479-488.
	LI Y F, WANG L, TANG C Y, et al. Design of smart public transportation facilities aided by AIGC technology[J]. Packaging Engineering, 2024, 45(6): 479-488. (in Chinese)
[18]	厉向东, 夏涵飞, 单逸飞, 等. 图形用户界面自动生成技术的现状与挑战[J]. 图学学报, 2024, 45(3): 409-421. DOI
	LI X D, XIA H F, SHAN Y F, et al. Opportunities and challenges: automatic generation technologies for graphical user interfaces[J]. Journal of Graphics, 2024, 45(3): 409-421. (in Chinese) DOI
[19]	冯玉泉. AIGC在工业设计上的应用与思考[J]. 包装工程, 2024, 45(8): 337-345.
	FENG Y Q. Application and thinking of AIGC in industrial design[J]. Packaging Engineering, 2024, 45(8): 337-345. (in Chinese)
[20]	黄喆, 王选政, 李杰. 人工智能与生成: 计算机介入产品设计[J]. 设计, 2024, 37(9): 77-80.
	HUANG Z, WANG X Z, LI J. Artificial intelligence and generation: computer intervention in product design[J]. Design, 2024, 37(9): 77-80. (in Chinese)
[21]	王茜. 基于AIGC背景下的产品设计流程与方法初探[J]. 鞋类工艺与设计, 2024, 4(4): 66-69.
	WANG Q. Study on product design process and method based on AIGC[J]. Shoes Technology and Design, 2024, 4(4): 66-69. (in Chinese)
[22]	于水, 范德志. 新一代人工智能(ChatGPT)的主要特征、社会风险及其治理路径[J]. 大连理工大学学报: 社会科学版, 2023, 44(5): 28-34.
	YU S, FAN D Z. The main characteristics, social risks and governance paths of the new generation of artificial intelligence (ChatGPT)[J]. Journal of Dalian University of Technology: Social Sciences, 2023, 44(5): 28-34. (in Chinese)
[23]	卢兆麟, 宋新衡, 金昱成. AIGC技术趋势下智能设计的现状与发展[J]. 包装工程, 2023, 44(24): 18-33.
	LU Z L, SONG X H, JIN Y C. State of arts and development of intelligent design methods under the AIGC trend[J]. Packaging Engineering, 2023, 44(24): 18-33. (in Chinese)
[24]	EDITORIALS N. Tools such as ChatGPT threaten transparent science; here are our ground rules for their use[J]. Nature, 2023, 613(7945): 612.
[25]	ALIŞIK E. “All compressed and rendered with a pathetic delicacy that astounds the eye”: Midjourney renders ambergris as constantinople[J]. CyberOrient, 2022, 16(2): 76-88.
[26]	杨莉莉. AI图像生成: 终结艺术契约[J]. 艺术当代, 2023, 22(3): 54-57.
	YANG L L. AI image generation: ending the art contract[J]. Art China, 2023, 22(3): 54-57. (in Chinese)
[27]	李雪瑞, 侯幸刚, 杨梅, 等. 基于多层次灰色综合评价法的工业设计方案优选决策模型及其应用[J]. 图学学报, 2021, 42(4): 670-679.
	LI X R, HOU X G, YANG M, et al. The optimal decision-making model of industrial design scheme based on multi-level grey comprehensive evaluation method and its application[J]. Journal of Graphics, 2021, 42(4): 670-679. (in Chinese)
[28]	SAATY R W. The analytic hierarchy process—what it is and how it is used[J]. Mathematical Modelling, 1987, 9(3/5): 161-176.
[29]	路鹏, 陈丽伶, 刘莹, 等. 产品设计实践教学的新路径——高校产品设计专业与护养院参与式设计的启示[J]. 高教学刊, 2019, 5(6): 95-98.
	LU P, CHEN L L, LIU Y, et al. New approaches in product design education: insights from participatory design between university product design programs and nursing homes[J]. Journal of Higher Education, 2019, 5(6): 95-98. (in Chinese)
[30]	HSIAO S W. Integrated FSM STM and DFA method to faucet design[J]. Journal of the Chinese Institute of Industrial Engineers, 1996, 13(3): 225-235.
[31]	陈士俊, 阎祥安, 谢庆森. 产品造型设计原理与方法[M]. 天津: 天津大学出版社, 1994: 3-8.
	CHEN S J, YAN X A, XIE Q S. Principles and method of product modeling design[M]. Tianjin: Tianjin University Press, 1994: 3-8. (in Chinese)
[32]	CHOU J R. A Kansei evaluation approach based on the technique of computing with words[J]. Advanced Engineering Informatics, 2016, 30(1): 1-15.
[33]	WU F, LU P, LIN Y C. Research on the influence of wheelsets on the visual imagery of city bicycles[J]. Sustainability, 2022, 14(5): 2762.
[34]	LU P, LI Z, LI J, et al. A consumer-oriented design thinking model for product design education[J]. Interactive Learning Environments, 2023, 31(8): 4753-4771.

内容	Midjourney	Stable Diffusion
生成模式	支持3种生成模式，包括“文本到图像”“图像到图像”“图像到文本”	支持3种生成模式，同Midjourney
核心技术	多种深度学习模型和算法，主要包括生成式对抗网络(GAN)、卷积神经网络(CNN)和变分自编码器(VAE)	扩散模型(diffusion model)
训练数据	大量高质量的图像数据，注重图像数据的艺术风格和设计元素	大量公开的图像数据集及其文本对，如LAION-5B，注重图像数据的广泛性和全面性
提示词	强调抽象的艺术风格，一般使用较短的提示词便可生成优于Stable Diffusion的高品质且符合预期的图像内容	强调详细的场景描述，通常需要复杂的正向提示词和反向提示词
图像控制	局部重绘和扩展图等功能可实现对图像的控制，但可控性逊色于Stable Diffusion	除了局部重绘和扩展图等功能，内置的ControlNet插件和自主训练的LoRA模型可实现对输出内容的精准控制，如能够实现二维线稿到三维渲染效果的快速转换，以及风格化迁移等
插件支持	支持少量插件	由于代码开源，支持各种插件
部署方式	部署在Discord云端平台	支持本地部署(离线)

内容	Midjourney	Stable Diffusion
生成模式	支持3种生成模式，包括“文本到图像”“图像到图像”“图像到文本”	支持3种生成模式，同Midjourney
核心技术	多种深度学习模型和算法，主要包括生成式对抗网络(GAN)、卷积神经网络(CNN)和变分自编码器(VAE)	扩散模型(diffusion model)
训练数据	大量高质量的图像数据，注重图像数据的艺术风格和设计元素	大量公开的图像数据集及其文本对，如LAION-5B，注重图像数据的广泛性和全面性
提示词	强调抽象的艺术风格，一般使用较短的提示词便可生成优于Stable Diffusion的高品质且符合预期的图像内容	强调详细的场景描述，通常需要复杂的正向提示词和反向提示词
图像控制	局部重绘和扩展图等功能可实现对图像的控制，但可控性逊色于Stable Diffusion	除了局部重绘和扩展图等功能，内置的ControlNet插件和自主训练的LoRA模型可实现对输出内容的精准控制，如能够实现二维线稿到三维渲染效果的快速转换，以及风格化迁移等
插件支持	支持少量插件	由于代码开源，支持各种插件
部署方式	部署在Discord云端平台	支持本地部署(离线)

评价尺度	含义
1	同等重要
3	略微重要
5	重要
7	非常重要
9	绝对重要
2,4,6,8	中间值

评价尺度	含义
1	同等重要
3	略微重要
5	重要
7	非常重要
9	绝对重要
2,4,6,8	中间值

阶数	R.I.
3	0.58
4	0.90
5	1.12
6	1.24
7	1.32
8	1.41
9	1.46
10	1.49