The Study Of Semantic Quantification For Material Texture Elements And The Optimization Methods

Nowadays product design emerges to focus on the need of users, while theusers’ requirement for products also increases rapidly, and people pays more attention to the details. The material texture, which prefers as an important product detail, is needed to be studied very carefully to make the right choice in the realproduct design process. This study is based on the theory of Kansei Engineer,using the psychological semantics and network model optimization method, to establish the plastic texture model which can analyze the subjective evaluation of users and output the corresponding objective material data accordingly.Aims at the Chinese natural linguistic of the texture image description for the plastic material, this study screens to get the reasonable word. And consider the features of plastic material texture imagery questions, this study chooses thegeometry model of semantic similarity as the basement for the imagery semanticquantization model. Then, through the use of multidimensional scaling method to build the vocabulary similarity space, using the semantic difference to conductthe original quantization of subjective evaluation, taking the pearson distance tocalculate the lexical similarity, targeting at the five representative word, this study builds the lexical similarity matrix. Finally, with the similarity matrix, studyquantifies the subjective evaluation of users based on the target words, and thusestablish the semantic quantization model.At the perspective of genetic algorithm, this study conducts the optimizationsearch in the five-dimensional parameter space of plastic material texture. Targeted at the user’s subjective evaluation, study conducts the solution fitting processin the five dimensions of the typical words. Then, study takes the5dimensiondata of the plastic texture, which corresponding to the best optimization result, as the output of the model. What’s more, this study analyzes the original questionfrom the perspective of multi-objective optimization, carries out the5sub-goalsfitting of the5typical words, outputs the overall objective data fo the plasticmaterial texture. Then, this study takes the data as the guidance for the plasticmaterial choice in the product design process.Uses8groups of validate data to test the performance of optimized model, takes the conjunction features of subjection and objection in the Kansei Engineer questioninto consideration, this study analyze the the output error of plastic texture parametermodel, integrated the relative error and the real world application of this model, verifiesfrom both perspectives of the material parameter and different material groups, provesthe model optimized gets the well performance in the users’ subjective evaluation datafitting, and the related output has the guidance ability for the material texture choice inthe real world design process.