Multiscale style transfer based on a Laplacian pyramid for traditional Chinese painting
This addresses the specific challenge of style transfer for traditional Chinese paintings, which is an incremental improvement over existing methods focused on western oil paintings.
The paper tackles the problem of unnatural stylization when applying existing style transfer methods to traditional Chinese paintings, which have plain colors and abstract objects, by proposing a multiscale method based on Laplacian pyramid decomposition that learns features at different scales. The result is the generation of appealing high-quality stylized images, as demonstrated through comparisons with state-of-the-art methods.
Style transfer is adopted to synthesize appealing stylized images that preserve the structure of a content image but carry the pattern of a style image. Many recently proposed style transfer methods use only western oil paintings as style images to achieve image stylization. As a result, unnatural messy artistic effects are produced in stylized images when using these methods to directly transfer the patterns of traditional Chinese paintings, which are composed of plain colors and abstract objects. Moreover, most of them work only at the original image scale and thus ignore multiscale image information during training. In this paper, we present a novel effective multiscale style transfer method based on Laplacian pyramid decomposition and reconstruction, which can transfer unique patterns of Chinese paintings by learning different image features at different scales. In the first stage, the holistic patterns are transferred at low resolution by adopting a Style Transfer Base Network. Then, the details of the content and style are gradually enhanced at higher resolutions by a Detail Enhancement Network with an edge information selection (EIS) module in the second stage. The effectiveness of our method is demonstrated through the generation of appealing high-quality stylization results and a comparison with some state-of-the-art style transfer methods. Datasets and codes are available at https://github.com/toby-katakuri/LP_StyleTransferNet.