Research On Emotional Product Design Based On Deep Learning Technology

In the field of design,emotional design has become an important research hotspot and development trend.This article attempts to study emotions with cutting-edge deep learning techniques.First,based on the Norman three-level theory,a three-level model of design features and emotions was proposed.The design emotions were interpreted from a more systematic perspective.The design features and types of design emotions at each level were analyzed.The correspondence between design features and emotions provides theoretical support for designers in the process of emotional design,transfers the study of emotions to the study of design features,provides a new perspective on research.The three-level model has thus become the theoretical basis for this article.After that,an emotional labeling experiment of product pictures was performed.First,the current mainstream design emotions words were obtained through the expert description of typical pictures,and the resulting words were classified into three levels based on a three-level model for analysis.Questionnaire was used to mark the emotions and design features of 2054 Red-dot and IF winning product pictures in different categories,and determine the emotional label of each picture and the reason for the emotions.The labeling explores the similarities and differences in the emotions of different types of products in various fields,and the distribution of emotions at various levels in product design.The qualitative and quantitative analysis of the emotions in current product design is conducted.Based on the experimental results,we had found the shortcomings of emotional design in application,and explored solutions to the problems found.The labeling experiment establishes the necessary training data set for deep learning of emotions.With learning technologies,this study performs deep learning on the labeled images.With sufficient material training,accurate learning results can be achieved.Through the theoretical research in the previous part of the paper,the correspondence between design features and emotions can be seen as an answer to the deep learning process from a design perspective.After deep learning,it can effectively awaken the silent massive network images without manually labeling emotions one by one,providing designers and the public with efficient design emotion picture searching,stimulating design creativity,promoting the formation of design inspiration,providing emotional judgment and scoring of design works,helping designers modify designs,and even generating designs,to promote the application of emotion in design.In this way,during and after deep learning,an exploratory study of the theory and practice of emotional design is achieved.