TY - GEN
T1 - Stroke controllable fast style transfer with adaptive receptive fields
AU - Jing, Yongcheng
AU - Liu, Yang
AU - Yang, Yezhou
AU - Feng, Zunlei
AU - Yu, Yizhou
AU - Tao, Dacheng
AU - Song, Mingli
N1 - Funding Information:
Acknowledgments. The first two authors contributed equally. Mingli Song is the corresponding author. This work is supported by National Key Research and Development Program (2016YFB1200203), National Natural Science Foundation of China (61572428, U1509206), Fundamental Research Funds for the Central Universities (2017FZA5014) and Key Research, Development Program of Zhejiang Province (2018C01004) and ARC FL-170100117, DP-180103424 of Australia.
Funding Information:
The first two authors contributed equally. Mingli Song is the corresponding author. This work is supported by National Key Research and Development Program (2016YFB1200203), National Natural Science Foundation of China (61572428, U1509206), Fundamental Research Funds for the Central Universities (2017FZA5014) and Key Research, Development Program of Zhejiang Province (2018C01004) and ARC FL-170100117, DP-180103424 of Australia.
Publisher Copyright:
© Springer Nature Switzerland AG 2018.
PY - 2018
Y1 - 2018
N2 - The Fast Style Transfer methods have been recently proposed to transfer a photograph to an artistic style in real-time. This task involves controlling the stroke size in the stylized results, which remains an open challenge. In this paper, we present a stroke controllable style transfer network that can achieve continuous and spatial stroke size control. By analyzing the factors that influence the stroke size, we propose to explicitly account for the receptive field and the style image scales. We propose a StrokePyramid module to endow the network with adaptive receptive fields, and two training strategies to achieve faster convergence and augment new stroke sizes upon a trained model respectively. By combining the proposed runtime control strategies, our network can achieve continuous changes in stroke sizes and produce distinct stroke sizes in different spatial regions within the same output image.
AB - The Fast Style Transfer methods have been recently proposed to transfer a photograph to an artistic style in real-time. This task involves controlling the stroke size in the stylized results, which remains an open challenge. In this paper, we present a stroke controllable style transfer network that can achieve continuous and spatial stroke size control. By analyzing the factors that influence the stroke size, we propose to explicitly account for the receptive field and the style image scales. We propose a StrokePyramid module to endow the network with adaptive receptive fields, and two training strategies to achieve faster convergence and augment new stroke sizes upon a trained model respectively. By combining the proposed runtime control strategies, our network can achieve continuous changes in stroke sizes and produce distinct stroke sizes in different spatial regions within the same output image.
KW - Adaptive receptive fields
KW - Neural Style Transfer
UR - http://www.scopus.com/inward/record.url?scp=85055574049&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85055574049&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-01261-8_15
DO - 10.1007/978-3-030-01261-8_15
M3 - Conference contribution
AN - SCOPUS:85055574049
SN - 9783030012601
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 244
EP - 260
BT - Computer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings
A2 - Ferrari, Vittorio
A2 - Sminchisescu, Cristian
A2 - Weiss, Yair
A2 - Hebert, Martial
PB - Springer Verlag
T2 - 15th European Conference on Computer Vision, ECCV 2018
Y2 - 8 September 2018 through 14 September 2018
ER -