Application and Research Analysis of Image Style Transfer Based on Neural Networks

University of Michigan, 500 S State Street, Ann Arbor, MI 48109, USA

^* Corresponding author’s email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract

Style transfer, as a key branch in the field of computer vision, is rapidly developing in various fields such as art creation and film and television. This article summarizes the image style transfer methods based on neural networks in recent years, and divides them into two types: GAN and diffusion models based on network structure. The variations of these methods with different structures are compared and analyzed, and the main characteristics of each category method and the suitable image categories for processing are summarized. Then, based on the application of style transfer in several commonly used fields, case studies were presented to demonstrate its value. Finally, challenges were identified from the perspectives of fidelity, cost, and ethical and legal aspects of style transfer, and corresponding solutions and research directions for the future were proposed. This article provides a comprehensive analysis of the development and application of image style transfer, hoping to provide research directions for researchers in related fields.