Research On Design Methods For Pattern Generation For Dunhuang Algal Well

Dunhuang caisson patterns are rigorous in configuration and contain rich variations in unity,which reflects ancient Chinese people’s aesthetic pursuit of simple decorative patterns.At present,the research of traditional pattern mostly focuses on the research of pattern content,and the redesign of pattern pattern is simple and lack of innovation.At the same time,the research based on the traditional pattern composition form is still insufficient.In view of the above problems,this paper takes the configuration of Dunhuang caisson pattern as the main body of research,and carries out the research on the pattern generation design pattern based on Dunhuang caisson pattern.The purpose is to get the method and process of pattern design generation by analyzing the pattern structure and form sense law of caisson.The main research contents of this paper are as follows:Firstly,combined with the theory of pattern composition,the structure form of caisson is analyzed from the overall frame and local form.Based on the clue of historical time,the evolution characteristics of caisson configuration are expounded in stages.Meanwhile,by referring to the relevant theories of Chinese and foreign pattern composition,the formal rules are analyzed and discussed from the two aspects of the overall configuration and the combination form between patterns of Dunhuang caisson pattern.Using the recursive operation thinking of shape grammar theory,this paper dissolves the composition form of the Dunhuang caisson pattern,and obtains the rules of dissolving caisson pattern.Then,according to the constraint rules of caisson configuration,the pattern recombination model is built and the pattern data set is constructed.Based on the visual cognitive characteristics of Gestalt psychology,the measurement method of the skeleton of Dunhuang caisson is defined,and the pattern of skeleton form is clustered.The frame and pattern were combined according to the constitutive frame and proportion rules of Dunhuang caisson pattern,and the pattern filling template was obtained.The design constraint rules were coded,the pattern automatic filling model was built,a large number of patterns were generated,and the pattern data set was composed.Finally,on the basis of the reconstituted pattern data set,this paper introduces the generative adversarial network model to learn the image features,and generates new images characterized by the Dunhuang caisson pattern for design optimization.The combination of database and generative deep learning technology provides more possibilities and innovation for pattern design and the construction of pattern data set also provides a certain data basis for the extension of pattern design.This paper aims to explore the artistic value of the sense of form of the pattern configuration of the caisson in Dunhuang,explore the new pattern design mode under the new technological change environment,and construct the pattern data set with the sense of form of the caisson pattern from the overall perspective of the pattern configuration based on the context of artificial intelligence.At the same time,it explores the process of creative design generation research of Dunhuang caisson pattern,and tries to make traditional pattern design pattern with regular configuration.It is expected to provide feasible research methods and referential research ideas for the redesign of traditional patterns and provide pattern data support for design and application.