Design And Implementation Of Web Application For Commodity Appearance Design Generation Based On Deep Neural Networks

Nowadays,the commodity in the consumer market is much too homogenous and in an oversupply,people are paying more and more attention to the goods with differentiated design.In last decade,with the overall popularity of e-commerce,personal brands has experienced a blowout and originality has become the key to increase sales.Both consumers and merchants are looking increasingly at the design and creativity of product.Using machine learning technology and python tools,we design and implement an product design drawing generative web application which can automatically generate a commodity design drawing according to the commodity edge drawing online.The core part of the web application is a image generative model which consists of two parts made by deep neural network:a generator and a discriminator,the learning process of the model is the repeated minimax two-player game between the generator and discriminator until the model has the ability to generate realistic commodity design drawing.We deploy the image generative model to the web application,so the user only needs to upload the commodity edge drawing to web server,and the commodity design generated by the web application can be obtained online..We experiment with workstation equipped with Ubuntu operation system and GPU,implement the model with the deep learning tool TensorFlow,and develop the web application with the python tool Django.There are UT Zappos5K shoes image dataset and Amazon handbag image dataset in our experiment.Based on the two datasets,we train a shoe design drawing and a handbag design drawing generative model.Deploy mature models to web applications that help users generate novel shoe and handbag designs.The image generative model can be trained using different commodity images,so the application can generate different commodity design drawings.Our web application for commodity design generation is able to effectively help people find new designs for their commodity,meanwhile,reduce the time cost and funding.