Review of Deep Learning-Based Style Transfer Research
Zhiyuan He, Kaixu Han, Yufei Li · 2021
Currently, style transfer solves a lot of tedious work and greatly improves efficiency and cost. With the continuous development of style transfer, there is a growing demand for using intelligence to solve the work of game rendering, animation production, advertising design, film production, and so on. Thus, we introduce two main types of current style transfer methods: image style transfer methods based on image iteration and image style transfer method based on model iteration. The basic principles and methods of the two methods are explained in detail, and their advantages and disadvantages are illustrated. Finally, the current application environment and development direction of style transfer based on deep learning are summarized.