A Structure-Aware Deep Learning Network for the Transfer of Chinese Landscape Painting Style

Der Lor Way*, Chang Hao Lo, Yu Hsien Wei, Zen Chung Shih

*此作品的通信作者

研究成果: Conference contribution同行評審

摘要

Recently, deep learning technology has made a breakthrough in computer vision, image processing, and other fields. Some researchers suggested neural style transfer method using a convolutional neural network (CNN). They established the correlation of features in a neural network to be treated as the style. However, their performance is unacceptable for Chinese landscape painting. According to the property of the Chinese landscape painting, this paper proposes a novel two stage style transfer method that imitates multiple styles of Chinese landscape painting based on deep learning. The structure of an input photo was simplified in the first stage. Then, the result of the first stage was transferred into the final stylized image in second stage. A generative adversarial network (GAN) is applied to train in each stage. Besides, a novel loss function was proposed to keep the shape of the content image. Finally, our method haves successfully imitated several styles of Chinese Landscape ink painting.

原文English
主出版物標題Culture and Computing - 11th International Conference, C and C 2023, Held as Part of the 25th HCI International Conference, HCII 2023, Proceedings
編輯Matthias Rauterberg
發行者Springer Science and Business Media Deutschland GmbH
頁面326-337
頁數12
ISBN(列印)9783031347313
DOIs
出版狀態Published - 2023
事件11th International Conference on Culture and Computing, C and C 2023, held as part of the 25th International Conference on Human-Computer Interaction, HCII 2023 - Copenhagen, 丹麥
持續時間: 23 7月 202328 7月 2023

出版系列

名字Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
14035 LNCS
ISSN(列印)0302-9743
ISSN(電子)1611-3349

Conference

Conference11th International Conference on Culture and Computing, C and C 2023, held as part of the 25th International Conference on Human-Computer Interaction, HCII 2023
國家/地區丹麥
城市Copenhagen
期間23/07/2328/07/23

指紋

深入研究「A Structure-Aware Deep Learning Network for the Transfer of Chinese Landscape Painting Style」主題。共同形成了獨特的指紋。

引用此