Abstract
This paper presents a content-aware style transfer algorithm for paintings and photos of similar content using pre-trained neural network, obtaining better results than the previous work. In addition, the numerical experiments show that the style pattern and the content information is not completely separated by neural network.