Product modeling design method based on graph neural network and fuzzy inference theory

The method of graph neural network fused with fuzzy inference theory is adopted to conduct in-depth research and analysis on the design of product styling, and designs a method to be used in practical design. Starting from the perspective of quantifying aesthetic preferences, it explores the indicators of objectively quantified preferences and the intelligent design method of product styling under the trend of emotionality and provides a reference experience for optimizing the industrial design process. The product side profile form is deconstructed into 25 sets of planar coordinates, which are used as input data for graph neural network fusion fuzzy inference theory, and the physiological indexes that can represent users' aesthetic preferences are used as output data to build the model and train the network. According to the analysis of the survey data, among the three types of samples of linear type, curved type, and combined type, the percentages of the pictures with the highest degree of preference in the evaluation group reached 68.52%, 75.53%, and 61.11%, respectively, and the weighted scores were higher than those of the control group.& COPY; 2023 THE AUTHOR. Published by Elsevier BV on behalf of Faculty of Engineering, Alexandria University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).