Research on Neural Style Transfer Algorithm

2019 2nd International Conference on Advanced Materials, Intelligent Manufacturing and Automation

Yao, Chengsi^1 ; Li, Yuanhao^1 ; Qi, Yali^1

Information Engineering, Beijing Institute of Graphic Communication School, Beijing, Beijing

102600, China^1

关键词: Convolutional neural network;    Development trends;    Industrial fields;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/569/5/052107/pdf
DOI  :  10.1088/1757-899X/569/5/052107

来源: IOP

【 摘 要 】

Researchers have successfully applied the convolutional neural network (CNN) to style transfer. Since then, Neural Style Transfer (NST) has received widespread attention in both scientific and industrial fields. Researchers in the field of machine vision are constantly proposing ways to optimize image style migration. This paper aims to summarize the history of style transfer before and after the rise of CNN, classify the existing classical and improved algorithms, and compare the results of some of them. Finally, after this study, we put forward some suggestions on the development trend of image style transfer.