Abstract: To accurately grasp the user's emotional image perception rules for product colors,and creatively generate product color design schemes that conform to users' emotional preferences,tan emotional design method for product colors based on deep learning was proposed.Semantic Difference (SD) and GoogLeNet model were used to construct a product color emotional image data set.Based on the product color emotional image data set and Conditional-Deep Convolution Generative Adversarial Network (C-DCGAN),the model of product color design scheme generation was established,and the product color design schemes was generated innovatively.The taxi color design was taken as an example to verify the effectiveness and applicability of the proposed method.

Key words: image, product colore motion design, GoogLeNet model, conditional-deep convolutional generative adversarial network

摘要： 为准确把握用户对于产品色彩的情感意象感知规律,创新性生成符合用户情感偏好的产品色彩设计方案,提出一种基于深度学习的产品色彩情感化设计方法。首先运用语义差异法和GoogLeNet模型构建产品色彩情感意象数据集;然后,基于产品色彩情感意象数据集与条件深度卷积生成对抗网络(C-DCGAN)搭建产品色彩设计方案生成模型,创新性生成产品色彩设计方案。最后,以出租车色彩设计为例,验证了所提方法的有效性与适用性。

关键词: 意象, 产品色彩情感化设计, GoogLeNet模型, 条件深度卷积生成对抗网络