色彩长期以来被认为是影响用户体验偏好的重要变量。智能化工业的发 展为视觉领域带来了强有力的辅助，智能化色彩搭配蓬勃发展。然而现有模型 存在对产品的针对性不足，网络架构较浅等缺陷，在色彩搭配精准度和美观度 上都仍存在改进空间。 为增强色彩搭配推荐的针对性，提高配色效率，本论文以速溶咖啡包装色 彩搭配设计为例，构建数据集，融合用户体验，输入去掉环境色的产品包装色 彩图像与标注的相关指数，将数据划分为训练集和验证集，调试网络函数和相 关参数，融合色彩美学和用户体验相关理论构建基于 GoogLeNet 的评价模型 并检验其准确度，进行模型训练，进而输出关于产品色彩包装与用户体验相关 的评价数据集为进一步研发生成配色方案的系统提供程序基础。 论文的研究聚焦于速溶咖啡包装的色彩搭配设计，通过结合深度学习网 络和用户体验指数，基于 GoogLeNet 模型构建产品色彩搭配用户体验评价模 型。从而为设计实践提供科学的指导。为有效实施针对性的产品色彩设搭配设 计实践提供参考。

Color choices influence user preferences. Technological innovations in the field aim to improve visual aesthetics by refining color combinations. Current approaches show room for improvement in precision, affecting the accuracy of color matching. This research contributes to advancing color matching techniques by utilizing instant coffee packaging as a case study. A specialized dataset is established, with distinct training and validation segments. Fine adjustments are made to the network operations based on GoogLeNet architecture. Rigorous testing and training are conducted to ensure precise evaluation of color packaging. This investigation employs deep learning networks to optimize the assessment of color combinations, thus guiding the creation of visually appealing product designs.