Deep permutation type design for Chinese characters

Abstract

With their intricate visual structure and cultural significance, Chinese characters present a unique realm for exploration in graphic design. However, the integration of graphic elements, such as two-dimensional primitives that are not limited to straight or curved strokes, into Chinese-type design relies mainly on the designer’s intuition. This paper investigates the application of deep permutation design methods to Chinese type design to explore the esthetic potential of graphic-based type design in a systematic and algorithmic manner. For this purpose, we propose a novel framework and present an implementation named ‘Permutype’, which operates in three phases: representation, generation, and evaluation. Through a series of visual experiments, we demonstrate how our framework promotes novel permutations and esthetic possibilities for Chinese characters. The methodology offers flexible graphics and exhaustive optimal scheme searches based on objective criteria and has the potential for extensive font libraries.

Keywords:

• Deep permutation design
• Chinese characters
• graphic design
• type design
• neural networks

Acknowledgements

We thank the creators of the fonts 'Gen Jyuu Gothic Medium' and 'Fang Zheng Kai Ti Simplified' for the visual representations in the figures demonstrating our methods and outputs. We appreciate the unwavering support and encouragement provided by Yachao Zhang and Ju Chen during this challenging period. We are grateful for the insights and suggestions from Chenghao Du, Tiaohao Lu, Shiying Luo, and Di Lan throughout this academic journey. Their contributions have been essential to the successful completion of this research project.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

De’en Chen

De’en Chen is a PhD candidate at the College of Design & Innovation at Tongji University. He is interested in design, art, and AI. Email: [email protected].

Kostas Terzidis

Kostas Terzidis is a professor at the College of Design & Innovation at Tongji University. Previously, he was an associate professor at Harvard University GSD (2003–2011) and an assistant professor at UCLA (1995–2003). He is the author of five books: Kids AID (Tongji: 2023), Permutation Design (Routledge: 2014), Algorithms for Visual Design (Wiley: 2009), Algorithmic Architecture (Architectural Press: 2006), and Expressive Form (Spon:2003). Email: [email protected].

Filippo Fabrocini

Filippo Fabrocini is a professor at the College of Design & Innovation at Tongji University, the director of the Tongji Sustainable AI Lab, and among the founders of the Tongji AI Art Lab. Filippo Fabrocini is also affiliated with the Italy National Research Council (Institute for Computing Applications). His main areas of interest are Machine Learning, Quantum Neural Networks, Ethical AI, and AI Art. Email: [email protected].