Kansei Engineering Design Process with Morphometrics and Generative AI: Car Headlight Case

Kansei Engineering is a field of study developed by Prof. Mitsuo Nagamachi at Hiroshima University's Faculty of Engineering in the early 1970s. It uses psychological methods to analyze human sensitivity and establish the relationship between specific design parameters (such as color, shape, and surface finish) and sensitivity for use in design. Shigekazu Ishihara has been working on a method to treat design elements as statistical quantities rather than qualitative variables in multivariate analysis. This paper introduces the application of Morphometrics, which treats geometrical form as a statistical quantity, to Kansei engineering. We conducted a study to investigate the impact of different car headlight shapes on people's Kansei responses. By analyzing the results, we determined the numerical changes to the headlight shape that would enhance a specific Kansei response. When creating particular images to study potential product designs, making many images with partial design candidates is desirable based on Kansei engineering.　Designers traditionally created manual drawings or photo collages, which were time-consuming and resulted in unnatural images, making it difficult to represent their proposed designs accurately. The design process we propose uses a generative AI, which allows designers to specify the area and rough shape of the part they want to study and then output a realistic product image with some natural embedding in a short period. This feature allows designers to concentrate on designing using Kansei engineering rather than putting extra effort into creating images.