Efficient Artistic Image Style Transfer with Large Language Model (LLM): A New Perspective
Bo Pan, YuKai Ke · 2023
With the development of intelligent information systems, image style transfer technology has been widely known. However, many image style transfer methods can only target one style, which is inefficient in application. In this study, the novel efficient artistic image style transfer with LLM is proposed. To be gin with the study, the related single style transfer models are reviewed to serve as the background. Then, the novel image demosaicing is designed to serve as the pre-processing for the complex images. As the core of the model, novel style transfer algorithm is proposed with the LLM. The novel neural network organization is designed and the core functions are optimized. Furthermore, to validate the performance, the visualzied style transfer test is conducted and the numerical simulation results on the efficiency is tested.