Year: 2024
Type: conference-paper
Abstract: Neural style transfer (NST) has evolved significantly in recent years. Yet, despite its rapid progress and advancement, existing NST methods either struggle to transfer aesthetic information from a style effectively or suffer from high computational costs and inefficiencies in feature disentanglement due to using pre-trained models. This work proposes a lightweight but effective model, AesFA---Aesthetic Feature-Aware NST. The primary idea is to decompose the image via its frequencies to better disentangle aesthetic styles from the reference image while training the entire model in an end-to-end manner to exclude pre-trained models at inference completely. To improve the network's ability to extract more distinct representations and further enhance the stylization quality, this work introduces a new aesthetic feature: contrastive loss. Extensive experiments and ablations show the approach not only outperforms recent NST methods in terms of stylization quality, but it also achieves faster inference. Codes are available at https://github.com/Sooyyoungg/AesFA.
Institutions: Seoul National University, Brookhaven National Laboratory
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
Funders: U.S. Department of Energy, National Energy Research Scientific Computing Center, National Research Foundation, Seoul National University, National Research Foundation of Korea, Ministry of Science and ICT, South Korea, Samsung, Office of Science, Institute for Information and Communications Technology Promotion, Advanced Scientific Computing Research
OpenAlex ID: W4393160304