A generative method and evaluation study of Nüshu patterns based on Shape Grammar and parametric fractals

Abstract

This study tackles the dual challenges of cultural feature loss and weak algorithmic controllability in the digital innovation of Intangible Cultural Heritage (ICH) by proposing a novel method that integrates Shape Grammar rules with parametric fractals for the generative design of Nüshu patterns. The methodology involved formalizing the compositional principles of typical patterns to construct a Shape Grammar rule set, followed by a rigorous comparative experiment evaluating four groups: original patterns (A), standard Iterated Function System (IFS) generation (B), AI generation (C), and the proposed hybrid model (D). A comprehensive evaluation combining objective metrics (fractal dimension, SSIM) and subjective expert assessments revealed the superior performance of the proposed method, which achieved the highest scores in cultural fit (4.42/5), visual aesthetics (4.48/5), and innovation appropriateness, while maintaining over 85 % of traditional aesthetic features and increasing structural complexity by 27 %. The results confirm the effectiveness of the “Shape Grammar + parametric fractal” model in achieving a balance between algorithmic innovation and cultural authenticity, offering a viable path for the living inheritance of ICH.