Statistical Shape Analysis of Headlights

In Kansei Engineering, sample shapes have been treated as categorical variables in nominal scale. In this paper, we describe the method for focusing on product shapes concerning customers' Kansei, to avoid confusing effects on Kansei caused by other properties such as sizes, orientations, colors, textures, and so on. We standardized the shapes of sample products and made them into statistical values with morphometric methods. This preprocessing of the shape let us build mathematical models of shape space and relations between shapes and Kansei based on the results of customers Kansei evaluations. Then, statistical analyses that are frequently used in Kansei Engineering, e.g., cluster analysis and principal component analysis, followed to reveal the characteristics of shape samples associated with specific Kansei and semantic space. Here, we described the proposed analysis method on headlights of passenger cars and present the resulting associations between concrete shapes and Kansei evaluation.