Aesthetic shape generation system based on novelty and complexity

Abstract

Increasing the aesthetic quality of a shape is an important objective for enhancing the attractiveness of a product. Many studies on shape generation systems are based on the user’s emotion as a means of supporting the ideas of product designers. However, most of the proposed systems are unsuitable for extrapolation of shapes because they rely on the subjective evaluation of the user to determine the shape of an object. Experimental aesthetics theory suggests that a moderate level of novelty and complexity yields pleasant feelings. Furthermore, a mathematical model has been proposed for this theory that focuses on the sum of the information content induced by the novelty and complexity of an object. In this study, we formulated the novelty and complexity of contoured shapes and developed a system that generates a variety of shapes with the given novelty and complexity parameters. We conducted experiments using the generated butterfly and automobile shapes. The results indicate that the system can independently manipulate the novelty and complexity of a shape and affect its interestingness. The system may support the evaluation and further investigation of the most acceptable level of novelty and complexity with respect to the product shape.

KEYWORDS:

• Aesthetics
• shape generation
• Kansei
• novelty
• complexity

Acknowledgements

The authors thank Dr Kazutaka Ueda, Ms Makoto Watarizaki and Prof. Tamotsu Murakami of The University of Tokyo for supporting this study. The authors also thank to the open source of PGGAN authors on GitHub.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are openly available in ‘ga_hybrid_gan_master’ at https://onl.bz/UZAUc45.

Additional information

Funding

This work was supported by JSPS KAKENHI grant number 21H03528.