A product form design method integrating Kansei engineering and diffusion model
Chaoxiang Yang,Fei Liu,Junnan Ye
DOI: https://doi.org/10.1016/j.aei.2023.102058
IF: 8.8
2023-07-05
Advanced Engineering Informatics
Abstract:The experience economy has shifted consumer demands from the functional to the emotional, and the emotional demands of the user have become a key design consideration. At the same time, product form design often relies solely on the knowledge and experience of designers, resulting in uneven design results and difficult quality assurance. To this end, this paper proposes a product form design method that applies an image generation algorithm oriented toward satisfying the emotional demands of users. Firstly, product images from the network are collected and processed to build a dataset of product images, which is used to train the Diffusion Model (DM) to generate product images that differ from the dataset. Secondly, the Kansei factors are obtained by clustering Kansei words from online reviews using Factor Analysis (FA) and then calculating the weights of the Kansei factors by the Analytic Hierarchy Process (AHP). Thirdly, a questionnaire is distributed to obtain user scores on the Kansei factors of the product images, and the Kansei evaluation value is calculated by weighting, then a prediction model is constructed using Support Vector Regression (SVR) to score and filter the generated images. Finally, the designer selects the highest-scoring images for detailing and tests the effectiveness of the design through user satisfaction. Using the ear thermometer as an example, we have created a product form that meets the emotional demands of users and verifies the scientific validity and effectiveness of the proposed method.
engineering, multidisciplinary,computer science, artificial intelligence