基于支持向量机回归与模拟退火算法的产品外观意象设计

Product appearance image design based on support vector machine regression and annealing algorithm simulation

针对产品造型、色彩和材质等外观特征要素与用户情感需求的复杂关联性问题,提出基于支持向量机回归和模拟退火算法的产品外观意象优化设计方法。结合聚类分析、因子分析等方法确定代表性样本和感性意象词汇对,借助语义差异法制作问卷,构建产品外观意象特征评价量表;通过支持向量机回归方法构建产品外观意象评价模型;将评价模型作为模拟退火算法的目标函数,优化产品外观意象设计,并建立产品外观设计推荐系统。以汽车方向盘为例进行设计验证,证明文中方法的有效性,为产品设计提供有效的辅助和支持。

Aiming at Complex relevance problems between appearance feature elements(including production modeling,color and materials)and user emotional requirement,an image optimization design method of product appearance was proposed based on support vector machine regression(SVMR)and simulated annealing algorithm(SAA).The representative samples and perceptual image word pairs were determined using cluster analysis and factor analysis.Questionnaires were made using semantic differential method to construct product appearance image feature evaluation scales.Support vector machine regression method was used to construct product appearance image evaluation model.The evaluation model was used as a target function of simulated annealing algorithm to optimize product appearance image design?and build product appearance design recommendation system.The method in this article was proved effectively with the example of automobile steering wheel.This could provide effective assist and support for production design.