首页 | 本学科首页   官方微博 | 高级检索  
     

基于深度学习的产品外观意象情感计算服务平台研究
引用本文:窦金花,覃京燕.基于深度学习的产品外观意象情感计算服务平台研究[J].包装工程,2020,41(6):20-25.
作者姓名:窦金花  覃京燕
作者单位:1.天津理工大学,天津 300384;2.北京科技大学,北京 100083,2.北京科技大学,北京 100083
摘    要:目的产品外观意象激发用户的情感体验,成为影响消费者购买决策的重要因素。情感的主观性与意象的模糊性,使产品外观意象与用户情感需求的有效映射成为产品设计的难题。产品外观具有多维特征,传统产品意象研究多考虑产品部分特征,需要人工标注产品特征,样本量小,造成设计方案与用户情感需求的偏差。针对这些问题提出解决方案,减少产品情感研究中不确定性因素的干扰,实现产品设计方案与用户意象需求的有效匹配。方法提出构建基于深度学习的产品外观意象情感计算服务平台,研究平台构建方法与关键技术。提出产品外观情感数据库建立方法和产品外观意象情感计算方法。构建产品外观意象情感计算模型,实现更为精确的产品外观意象情感分类与情感特征识别。提出分析不同产品类型的产品外观情感特征权重,研究产品特征对情感意象的贡献程度,挖掘关键情感特征。研究面向用户情感需求的产品外观方案检索方法和技术,以构建产品设计服务平台。结论研究成果将为建立高效、人性化的产品外观意象设计协同服务平台提供参考,从而简化设计师与用户的知识获取过程,促进产品外观设计方案与用户意象需求的有效匹配。

关 键 词:人工智能  创新设计  深度学习  情感计算  产品外观意象
收稿时间:2020/1/20 0:00:00
修稿时间:2020/3/20 0:00:00

Affective Computing Service Platform of Product Appearance Image Based on Deep Learning
DOU Jin-hua and QIN Jing-yan.Affective Computing Service Platform of Product Appearance Image Based on Deep Learning[J].Packaging Engineering,2020,41(6):20-25.
Authors:DOU Jin-hua and QIN Jing-yan
Affiliation:1.Tianjin University of Technology, Tianjin 300384, China; 2.University of Science and Technology of Beijing, Beijing 100083, China and 2.University of Science and Technology of Beijing, Beijing 100083, China
Abstract:The product appearance image stimulates the user’s emotional experience,which becomes a key factor to influence the consumer’s purchasing decision.Considering the subjectivity of emotion and the ambiguity of image,how to realize the effective mapping of product appearance image and emotional needs becomes a difficult problem of product design.There are multi-dimensional features of product appearance.The traditional product image research often considers the partial features of the product and needs to manually mark the product features,and the sample size is small.These are likely to cause the design scheme to deviate from user’s emotional needs.In response to these questions,the work aims to propose solutions which can reduce the interference of uncertainty factor in product’s emotional research,and achieve the effective matching of product design schemes and user image requirements.A proposal was made to build a deep learning-based affective computing service platform of product appearance image to study the construction methods and key technologies of the platform.The establishment method of product appearance emotional database and the affective computing methods of product appearance image were proposed.The affective computing models of product appearance image were constructed to realize the emotion classification and emotion feature recognition more precisely.A proposal was made to analyze the emotion feature weight of the product appearance for different product types,and study the contribution of product features to emotional image to dig key emotion features.The retrieval methods and technologies of product appearance schemes oriented to users’emotional needs were studied to build a product design service platform.The research results will provide the reference for the establishment of an efficient and humane design collaborative service platform of product appearance image,thereby simplifying the knowledge acquisition process for designers and users,and promoting the effective matching of product appearance design schemes and user image requirements.
Keywords:artificial intelligence  innovation design  deep learning  affective computing  product appearance image
本文献已被 CNKI 维普 等数据库收录!
点击此处可从《包装工程》浏览原始摘要信息
点击此处可从《包装工程》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号