Research on redesign of Chinese local embroidery patterns based on extension semantics and the pix2pix model

Abstract

In order to solve problems such as the cultural connotation and stylistic characteristics of the pattern ontology that are easily ignored in the design of traditional Chinese local embroidery patterns, in this paper, a regeneration design method of Chinese local embroidery pattern is proposed based on extensible semantics and the pix2pix model. First, the extension semantic analysis method is used to summarize and quantify the embroidery pattern of Hong’an, and the feature words are illustrated with the help of graphic thinking to obtain the optimal extension interval diagram as a design reference. Then, the data set of the Hong’an embroidery pattern is created by the sample collection method, and the preprocessed data set is input into the pix2pix algorithm for training to obtain the optimal number of iterations and output high-quality design patterns. Finally, we implement a subjective–objective evaluation framework to assess five key indices: cultural expressiveness, visual aesthetics, graphic innovativeness, peak signal-to-noise ratio, and structural similarity of the generated design patterns, and the design scheme with high merit is selected to carry out design practice. The experimental results show that our method is effective in realizing various satisfying vivid regeneration designs of Hong’an embroidery patterns and further providing new perspectives for the inheritance and development of intangible cultural heritage.