Research on the Construction System of Digital Art Creation Platform Based on Deep Learning

This paper presents an algorithm for creating art style based on deep convolutional neural networks. This platform can automatically transform artistic style and create creative content by means of deep learning algorithm. It also supports user input image's artistic style. To test the validity of the platform, the paper has designed an experiment and used a variety of art data sets to train and evaluate the model. Experiments indicate that this platform can be used to transform the input images into different artistic styles. The produced art works are in the same way as the traditional artistic creation, but also display the great variety of creation. The generated images not only meet the visual characteristics of the art style, but also maintain the content integrity of the input image. The experiment results show that this platform is better than the traditional one in generating speed and creating effect, and the generation error is less than 5%, which shows that this platform is very useful in artistic creation.