基于深度学习的情感化设计

发布时间：2018-07-13 17:44

【摘要】：目的使用深度学习实现情感化设计,满足用户个性化的情感需求,加速传统设计过程,提升用户体验。方法研究深度学习中可用于情感化设计的算法,使用卷积神经网络(CNN)实现名画复制品的个性化自动生成;抓取互联网数据,使用LSTM模型挖掘用户真实需求,进行当前流行游戏的周边产品设计;以张裕葡萄酒庄旅游纪念品设计为例,使用深度学习基于用户个人信息和行为数据推荐个性化的葡萄酒包装。结论基于CNN的名画复制品的个性化生成丰富了图像的可修改空间,满足了用户个性化的情感诉求;基于LSTM的用户需求分析高效和准确地反映了用户的真实需求,加速了传统用户调研过程;基于深度学习的旅游纪念品个性化设计进一步提升了用户体验。将深度学习应用于情感化设计有利于挖掘用户内心的真实需求,节省人力物力,满足用户情感诉求和提升用户体验,进一步为设计学与计算机科学的交叉提供了有效方法。
[Abstract]:Objective to realize the emotional design with deep learning, to meet the individual emotional needs of users, to accelerate the traditional design process and to improve the user experience. Methods the algorithm for emotional design in deep learning was studied. The personalized automatic generation of famous painting replicas was realized by using convolution neural network (CNN), and the real needs of users were mined by using LSTM model, grabbing the Internet data and using LSTM model. Design the peripheral products of the current popular game, take Changyu winery tourism souvenir design as an example, use in-depth learning to recommend personalized wine packaging based on user's personal information and behavior data. Conclusion the personalized generation of famous painting replicas based on CNN enriches the image modifiable space and satisfies the emotional appeal of users, and the analysis of user needs based on LSTM reflects the real needs of users efficiently and accurately. The traditional user research process is accelerated, and the personalized design of tourist souvenirs based on deep learning further improves the user experience. The application of deep learning to emotional design is helpful to excavate the real needs of users, save manpower and material resources, satisfy the emotional demands of users and improve user experience, which provides an effective method for the intersection of design and computer science.
【作者单位】：北京科技大学;
【基金】：国家自然科学基金(61602033) 北京市科技计划项目(Z171100001217009) 北京市社会科学基金(16YTC027) 中央高校基本科研业务费资助(FRF-TP-15-027A1)
【分类号】：TB472

【相似文献】