The Study Of Quantitative Relationship Between Product Material Texture Image And User Preferences

Material as the material basis of the product, the material texture image cognition largely around the user for the choice of products, the research of texture image will help according to user requirements for product design. Based on the commonly used material as the research object, discuss the relationship between the material texture image and user preferences, and set up the material texture image with the user preference quantitative relationship model, referred to the quantitative model for texture image, and through the model analysis the quantitative relationship between material texture image and user preferences.First of all, through the full investigation and analysis, summarized the main influence parameters of various materials texture changes; From original features of material, on the basis of the objective physical parameters of metal, plastic, glass, wood material,use visual test to determine the scope and the parameters of material samples. Through the comparative analysis in order to receive the word can accurately describe the material texture image vocabulary; The word similarity test and multidimensional scaling analysis will gather in class image word group, get the representation of the words in each group; Using semantic differential method to quantify vocabulary typical image of preference of material samples.Secondly, according to the nonlinear relations of material parameters and image quantitative values, using the BP neural network to discuss quantitative relations between the material types, roughness and refractive index material parameters and representative image vocabulary. Through material texture quantitative test data training BP neural network, obtain the quantitative model of the relationship between the material parameters and the image quantization value, through the verification the model is realized to predict the degree of preference of the representative images based on the material parameters. According to model to analyze the quantitative relationship different material of parameter changes the impact on the emotional user preferences.Finally, quantified the user perceptual demand, though the word similarity experiment, transform quantitative results into quantitative data of nine representative image vocabulary. Using genetic algorithm and improved parallel choice method comparison and analysis, summarize the problem into many image target problem, the material parameters of the optimal results are obtained and greatly improve the fitting accuracy. Through the data validation, prove that the model has good fitting performance for image of user preference value, the optimal solution is the best material parameters can meet the demand of user perception. The model has the reference value and the instruction significance for the designer according to the user’s preference demand choice material.