Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : September 20, 2022 - September 22, 2022
Design aesthetics is one of the most important factors in the attractiveness of industrial products. In this study, we focused on the theory of aesthetics, which states that moderate novelty and complexity maximize pleasant feelings, and formulated the novelty and complexity perceived from a closed curve contour shape using curvature and Fourier expansion coefficients. Using this formulation, we developed "Hybrid-GAN," a shape generation system that iterates shape generation, feature computation, and shape optimization. The system is capable of generating a variety of shapes of arbitrary novelty and complexity. In a subsequent experiment, subjective evaluations of novelty and complexity, as well as beauty and interest, were obtained for the shape samples generated by the system. The results showed that the novelty and complexity formulations had significant positive correlations with the subjective evaluation, and the sum of novelty and complexity had a strong positive correlation with interestingness. The results of this study are expected to be used to support the design of attractive and long-loved shapes by providing feedback to designers on the degree of novelty and complexity that users find most pleasant.
