Image Artistic Style Migration Based on Convolutional Neural Network

In recent years, the wave of artificial intelligence technology, which is guided by deep learning, is becoming more and more widely applied to all fields of society. Among them, the cross collision between artificial intelligence and art has attracted great attention in related research fields. The migration of image artistic style based on deep learning has become one of the active research topics. In this paper, a simple and effective method is presented for image artistic style migration. That is, firstly, we specify an input image as an original image (it is also called a content image); at the same time, another or more images are designated as the desired image style. And then, by constructing the network model based on convolutional neural network (CNN), the image style can be transformed while the content information of the content image is guaranteed, so that the final output image shows the perfect combination of the content of the input image and the style of the style image. The core of the proposed artistic style migration strategy is the construction of an unified CNN framework. Here, a generation network is set up based on a deep residual network and the VGG-19 network model is applied to built a loss network. The experimental results on an application system show that our proposed method achieves a good synthesis effect for image artistic style migration.