基于人工神经网络的电动汽车前大灯意象造型设计研究

Image Modeling Design of Electric Vehicle Headlight Based on Artificial Neural Network

中文关键词: 电动汽车造型  前大灯造型  人工神经网络  感性意象

英文关键词: electric vehicle modeling  headlight design  artificial neural network  perceptual image

中文摘要:

      目的 针对电动汽车造型设计中的前大灯意象造型问题，分析设计要素与用户情感之间的关系。方法 分析现有电动汽车前大灯样本，提出一种新的前大灯造型描述方法，形成设计要素。结合用户意象调研分析方法获得关键意象，建立语义量表;由设计要素与语义量表构成训练样本，建立神经网络模型;基于神经网络的泛化能力，结合了枚举算法与控制变量思想，分析设计要素与用户情感的关联性。结论 分析获得目标意象下的电动汽车前大灯设计要素最优组合形式，以及设计要素与目标意象的相关性;结合实验数据提出造型综合评价模型，将用户对前大灯造型美观程度的感知解释为对其造型的综合评价，以设计要素为线索，造型综合评价与目标意象评价呈源与流的关系;基于造型综合评价模型中设计要素、综合评价、目标意象评价之间的联系，通过分析设计要素与目标意象评价的相关性，间接获得设计要素与造型综合评价的相关性，作为设计活动中保证预期造型综合评价和各目标意象评价的参考。

英文摘要:

      Aiming at the problem of headlight image in electric vehicle modeling design, the relationship between design elements and user emotion is analyzed. The existing headlight samples of electric vehicles are analyzed, and a new headlight modeling description method is proposed to form design elements. Combined with the research and analysis method of user image, the key image is obtained and the semantic scale is established; the training samples were composed of design elements and semantic scale, and the neural network model was established; based on the generalization ability of neural network, combining enumeration algorithm and control variable idea, the correlation between design elements and user emotion is analyzed.The optimal combination form of design elements and the correlation between design elements and target image are analyzed. According to the experimental data, a modeling comprehensive evaluation model is proposed to comprehensively evaluate the modeling by explaining the user’s perception of the beauty of the headlight modeling. Taking the design elements as the clue, the relationship between the comprehensive evaluation of the modeling and the evaluation of the target image is in the form of source and flow. Based on the relationship among design elements, comprehensive evaluation and target image evaluation, the correlation between design elements and comprehensive evaluation of modeling is indirectly obtained by analyzing the correlation between design elements and target image evaluation, which can be used as a reference to ensure the comprehensive evaluation of expected modeling and each target image evaluation in design activities.