Visual emotional resource extraction and communication design based on improved style transfer algorithm

This article introduces a study on the design and dissemination of visual emotional resources through an improved style transfer algorithm. With the advent of the digital communication era, visual emotional resources have become an indispensable factor in cultural and artistic creation, as well as in commercial advertising and marketing. Therefore, this study first presents a new style transfer algorithm based on the style-embedded network DaseNet. The algorithm uses the convolutional layers of a convolutional neural network (CNN) network to fuse the content image and the style image. Different feature style-embedded networks are combined on 3 × 3 convolutional layers, and the reduced style features are input into the decoder to obtain the stylized image. The improved style transfer algorithm (ISTA) has fast iterative optimization and is suitable for transferring more diverse styles. In the dissemination design of visual emotional resources, the article provides data support by selecting appropriate social media and interactive modes. The research in this article provides new ideas and methods for the design and dissemination of visual emotional resources extraction. It is expected to have a positive impact on cultural and artistic creation, as well as commercial advertising and marketing in the digital communication era.